Privacy from the Government

Readings:The next segment we will do is on Privacy. Start reading Baase Chapter 2.

Privacy and the government
Nothing to Hide
NSA Surveillance
Parallel Construction
Microsoft v US
Supreme Court Cases on Privacy
    Third-party doctrine
Electronic Communications Privacy Act

Opinions on Internet Privacy

They are watching you: The speaker is Jerry Day.

Is this a real threat? (See especially the section between 0:45 and 1:25)

I'm offering this as an example of a possible threat, but which definitely has elements of "paranoia" as well. (I imagine somewhere on YouTube there's a video of someone explaining the dangers of the government eavesdropping on your conversations by beaming lasers on your windows.)

That said, the organization Naperville Smart Meter Awareness sued the city of Naperville, IL, starting in 2011. The city of Naperville is the utility for Naperville residents, and they were collecting electrical-usage data every 15 minutes. The Seventh Circuit ruled that this data collection was in fact a "search", but that in this particular case it didn't matter as the data was not being used as evidence. See See also Jonas Diener.


What is privacy all about? Baase (4e p 48 / 5e p 52) says it consists of
Are these all? Note that Baase put control of information as #2; I moved it to #1.

In some sense the second one above is really a different category: the need to get away from others. A technological issue here is the prevalence of phones, blackberries, and computers and the difficulty of getting away from work. (There is also a touch of class privilege here: lower-income people have a much-diminished expectation of having access to any private space (including the bathroom).)

The third one is to some degree a subset of the first: who gathers information about us, and how is it shared? Another aspect of the third one is freedom from governmental spying. Privacy from the government is a major part of Civil Liberties. When we talk about government surveillance, we often think of real-time surveillance: listening in on our phone calls, tracking our location, or reading our emails. However, a lot of government surveillance takes the form of retrospective searches of databases gathered a while ago: checking what organizations we belong to, where our car was seen a month ago, what purchases we made recently, or what websites we visited.

Privacy is largely about our sense of control of who knows what about us. We willingly put information onto Facebook, and are alarmed only when someone reads it who we did not anticipate.

Privacy from:

What do we have to hide

Sometimes, when we try to argue for our privacy from the government, we get asked what do you have to hide? See below. (The internet-advertiser version of this is often we're not collecting anything important).

On the one hand, many people who don't have anything to hide are nonetheless uncomfortable with surveillance. On the other hand, a big part of post-9/11 surveillance (eg, what the NSA is doing) is for protection of the public.

On yet another hand, should we care at all about privacy? Or is it just irrelevant?

MIT's Sherry Turkle gave the keynote talk at Loyola's Digital Ethics conference in Fall 2012. She shared some comments made by teenagers in privacy discussion groups:

Do we now have to be good all the time?

She also quotes Facebook founder Mark Zuckerberg as saying "privacy is no longer a relevant social norm".

Turkle's own rule for healthy democracies is that the government should assume "everyone has something to hide". That way, there can be no arguing that if you're in favor of greater privacy then you have "something to hide". But what if safety is involved? Turkle's approach is especially complicated when balancing a relatively large invasion of privacy against a relatively low risk.

Strange history: once upon a time we were mostly concerned about privacy from the government, not from private commercial interests. Then things shifted 180°; commercial interests were the primary concern.

Now they have shifted 180° again.

Once upon a time, concern about privacy was on the decline. People knew about the junk-mail lists that marketers kept, but it did not seem important, especially to younger people.

In the last few years, commercial privacy has become a significant issue. Why is this?

Psychologists have ways of defining general personality traits, eg the OCEAN set of
(The Myers-Briggs system has four dimensions, and classifies you as at one end or the other (eg extraverted or introverted) on each axis.)

We have reached the point that outsiders can create a psychological profile of us using online data only. Once upon a time, the potential for this was seen as frightening. But is it? Does it even matter if advertisers, or the government, know we score relatively low on Conscientiousness?

Is this even what we mean by losing our privacy? Psychologists have suggested that "getting to know someone" is based significantly on the slow voluntary exchange of personal information, which would include our personality traits.

Alternatively, maybe "losing our privacy" has meaning only when we have to confront the loss in immediate social settings. Perhaps the marketing information about us was too remote for us to be concerned. However, now that Facebook has ushered in a new era of online information that is indeed about our immediate social situation -- friends, events, likes -- maybe we feel the loss of privacy much more keenly.

In 2010, Tal Yarkoni did a study Personality and Blogging, in which he identified correlations between language use and traditional OCEAN-based psychological categories, using subjects who had consented to a standard psychological-profile evaluation. He was able to create a mechanism for determining someone's psychological categorization just from the language the person used in blogging.

In 2013, IBM's Eben Haber extended this to (much-shorter) twitter postings. The goal is indeed to make use of the inferred information about personality to target marketing efforts more effectively. The original (short) paper is PersonalityViz: A Visualization Tool to Analyze People's Personality with Social Media, by Liang Gou, Jalal Mahmud, Eben M. Haber and Michelle X. Zhou. See also

An important corollary of Yarkoni and Haber's work is that it appears to be much harder to conceal ones fundamental personality when online than some have perhaps thought.

In January 2015, Youyou, Kosinski and Stillwell published a paper in which they showed that Facebook likes also revealed ones OCEAN profile, and furthermore did this more accurately than family and friends: Computer-based personality judgments are more accurate than those made by humans. This research has been widely covered in the popular press: see, where coauthor Kosinski notes "one of our most surprising findings is that we could even predict whether your parents were divorced or not, based on your Facebook likes." You can also go to where you can upload your Facebook likes and then click "predict my profile".

And then there are all those Facebook quizzes. Many of these represent someone very deliberately mining for information about your OCEAN profile, to help determine the kinds of ads you'll see. This works even if Facebook keeps all the results: an advertiser can say "show this ad to everyone whose best-match dog breed is a pit bull and whose best-match flower is a thistle". See

Personality identification in the advertising world continues to grow more and more precise. Ordinary advertisers often have access to this information; Facebook has your personality down very precisely. Do you care? Yes, it can be used to help target ads to you, but only in conjunction with other information about your interests.


We understand that all sorts of online purchasing information is collected about us in order for the stores to sell to us again. Whenever I go to, I am greeted with book suggestions based on past purchases. But at what point does this information cross the line to become "personalized pitches"?

What if the seller has determined that we are in the category "price-sensitive shopper", and they then call/mail/email us with pitches that offer us the "best price" or "best value"? (See the box on Baase, 4e p 72 / 5e p 65, for a related example. Here, the British Tesco chain determined which shoppers were "price-conscious", and also what they were most likely to buy. These products (maybe the top 20 in sales volume?) were then priced below Wal*Mart's prices.)

Political parties do this kind of personalization all the time: they tailor their pre-election canvassing to bring up what they believe are the hot-button issues for you personally.

Marketing personalization sometimes involves your personality profile, but often weighs other attributes (like "price-sensitive shopper) more highly.

What do computers have to do with privacy?

Old reason: they make it possible to store (and share) so much more data
Newer reasons: 
Baase, 4e p 48 / 5e p 52: The communist East-German secret police Stasi were masters of non-computerized privacy invasion. The film The Lives of Others was about this.

The Fourth Amendment states:

The right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures, shall not be violated, and no Warrants shall issue, but upon probable cause, supported by Oath or affirmation, and particularly describing the place to be searched, and the persons or things to be seized.

Should this extend to any online information?

Note the requirement that the person and place be specifically described. Does this rule out broad online searches? (See "geofence" warrants, below.)

Baase 4e p 50 / 5e p 54: many companies use computers to create "a detailed picture of the person's interests, opinions, relationships, habits and activities".

Maybe also of what sales pitches we're likely to respond to?
Some non-governmental privacy issues:

Caller ID

When it first came out in the early 1990's, Caller ID was widely seen as a privacy intrusion. That is, it took away your "right" to call someone anonymously. Actually, that is a plausible right if you're calling a commercial enterprise; if you don't want them calling you back, you should be able to refuse to give them your number.

Within a decade, Caller ID was widely seen as a privacy boost: you could control who could interrupt you. This is privacy in sense #2 above; the original issue was privacy in sense #1.

Caller ID never caught on with stores; it did catch on with ordinary people.

Is there any right to phone someone anonymously? What if you're trying to give the police a tip? What if you're a parole officer?

Maybe some of the most sensitive information gathered about us today is our location, typically from a cellphone. Traditional phones do not necessarily track GPS in real time, unless an emergency call is placed, but smartphones do this continuously in order to display advertisements for nearby businesses. What undesirable things could be done with this information?

We will return to this later., listing twitter/foursquare announcements that you will not be At Home (now "off"; I wish I'd kept some sample data)
Facebook has made us our own worst privacy leakers.

Facebook and college admissions, employment, any mixed recreational & professional use

Here is a list of some specific things we may want to keep private, and which might also appear in records somewhere:
Of course, a central issue in the last item is what constitutes "minor".

In keeping these sorts of things private, are we hiding something?

More significantly, what has the rise of Facebook done to this list? How much do we care about this "general background" information as opposed to the kind of information that leaks out of Facebook: who we partied with last night, what we drank, who we partied with five years ago, where we were last night given that we said we would be volunteering at the soup kitchen?

Consider the item above about "different facades in different settings". In this context the following quote from Mark Zuckerberg is relevant:

The days of you having a different image for your work friends or co-workers and for the other people you know are probably coming to an end pretty quickly. ... Having two identities for yourself is an example of a lack of integrity" [from David Kirkpatrick, The Facebook Effect]

Lack of integrity? Really? The only thing that keeps LinkedIn alive is that most people believe in keeping at least some separation between their work life and pictures of their partying. But the separation goes much deeper than that; many people maintain different images in different contexts. See also

In Japan, there are terms for ones "true feelings" and ones "public opinions": honne and tatemae ( Divergence between the two is widely accepted, and is not regarded as "hypocrisy".

Sometimes we want to keep things private simply to avoid having someone else misinterpret them.

Is this list what is really important to us in terms of privacy? Or are we really only concerned with more intangible attributes?

In 1979, Harold Sackeim and Ruben Gur studied self-deception. They asked participants the questions below, with the understanding that the honest answer was almost always "yes" (this is debatable, but they do have a point). The number of "no" answers was then interpreted as an indication of self-deception.

For our purposes, the issue is that these questions represent another list of things we might very well wish to keep private (note that the survey dates from 1979, when taboos against homosexuality were still strong).

1. Have you ever felt hatred toward either of your parents?
2. Do you ever feel guilty?
3. Does every attractive person of the opposite sex turn you on?
4. Have you ever felt like you wanted to kill somebody?
5. Do you ever get angry?
6. Do you ever have thoughts that you don't want other people to know that you have?
7. Do you ever feel attracted to people of the same sex?
8. Have you ever made a fool of yourself?
9. Are there things in your life that make you feel unhappy?
10. Is it important to you that other people think highly of you?
11. Would you like to know what other people think of you?
12. Were your parents ever mean to you?
13. Do you have any bad memories?
14. Have you ever thought that your parents hated you?
15. Do you have sexual fantasies?
16. Have you ever been uncertain as to whether or not you are homosexual?
17. Have you ever doubted your sexual adequacy?
18. Have you ever enjoyed your bowel movements?
19. Have you ever wanted to rape or be raped by someone?
20. Have you ever thought of committing suicide in order to get back at someone?

For many of these, however, there are not any records (except for #8, if your friends' cameras were handy at the time).

Some data collection that we might not even be aware of:
Where do we draw the line? Or is there no line? Is loss of privacy a matter of "death by a thousand cuts"?

Privacy from the government

This tends not to be quite as much a computing issue, though facial recognition might be an exception. "Matching" (linking the names, say, of everyone receiving welfare payments and also owning a car worth more than $15K) was an example once upon a time. Interception of electronic communications generally fits into this category; the government has tried hard to make sure that new modes of communication do not receive the same protections as older modes. They have not been entirely successful.

One of the biggest issues with government data collection is whether the government can collect data on everyone, or whether they must have some degree of "probable cause" to begin data collection. On 4e p 69 / 5e p 87 of Baase there is a paragraph about how the California Department of Transportation photographed vehicles in a certain area and then looked up the registered owners and asked them to participate in a survey on highway development in that area.

Why might that have been considered to be a problem?

The California episode probably happened in the late 1990's. Does that matter?

Police departments (and their civilian contractors) across the US are now routinely scanning all license-plate numbers.

Canadian position: government must have a "demonstrable need for each piece of personal information collected".

Nothing to Hide

Why do we care about privacy? Is it true that we would not care if we had nothing to hide? What about those "minor transgressions"  on the list? Are they really minor?

Or is is true that, as Julian Sanchez wrote, "we live 'in a nation whose reams of regulations make almost everyone guilty of some violation at some point'" [Baase 4e p 63 / 5e p 84]

The "nothing to hide" question is central to privacy. But note the hidden assumption that you only need privacy if you do have something to hide!

Once upon a time (in the 1970's) there was some social (and judicial) consensus that private marijuana use was modestly protected: police had to have some specific evidence that you were lighting up, before they could investigate. Now, police are much more free to use aggressive tactics (eg drug-sniffing dogs without a warrant, though they can't use thermal imaging without a warrant).

Is this a privacy issue?

Now the NSA collects everyone's phone records, and sometimes (it is not entirely clear how often) uses the information to identify drug dealers (including marijuana dealers). The information may then be turned over to the DEA.

Is personal marijuana use an example of the kind of thing we have a "right" to keep hidden from the government? Or should the government make use of every possible tool to prevent this?

What about speeding?

What about claiming as a tax deduction a lunch with a colleague, during which you supposedly discussed business, but your pre-lunch texts to one another make it clear that you both really wanted to discuss a soccer match?

Perhaps "you should have nothing to hide" is a bit harsh. Maybe another way to phrase this is to say that, in the interests of preventing terrorism, child abuse, narcotics trafficking, organized crime and cybercrime, we should all cooperate to give law enforcement better access. What things might someone want to keep secret from the government?

Here are a few "Nothing to Hide" essays:

There is also Daniel Solove's paper "'I've got nothing to hide', and other misunderstandings of privacy",

Do you believe these arguments?

A more specific argument is the basis of Harvey Silverglate's book "Three Felonies A Day: How the Feds Target the Innocent". Unfortunately, many of Silverglate's examples relate to disclosure or non-disclosure or corporate-malfeasance issues that are very complicated. However, back (last year) when violating any of Loyola's computer policies was a felony, I might have had that issue. (For example, Loyola's policies required that servers only be used in server rooms, accessible only to ITS employees, and also that every laptop was a server (because svchost processes are servers!).)

How about end-to-end encryption of messages? That's a very specific thing that governments would like to put a stop to. Should we help them? What are some justifications for choosing end-to-end encryption of our messages?

The post-9/11 loss of American privacy to government surveillance was often justified by "if you have nothing to hide, you have nothing to fear". Ironically, the strongest proponents of this approach were on the political right, who are -- in 2021 -- rather obsessed with "personal freedom" and the right not to wear a mask.

A much more specific version of the "nothing to hide" argument is that having the NSA collect your phone metadata is a small price to pay for greater American safety in a post-9/11 world. Does your phone metadata implicate you in any way? (It might if you are leaking government secrets, or if you are in an illicit relationship.)

"Everyone is guilty of something or has something to conceal. All one has to do is look hard enough to find what it is" -- Alexandr Solzhenitsyn (the opinion of the character Rusanov -- a records manager -- in Cancer Ward)

There is also another approach to the "what do you have to hide" question: do you trust the government? You might feel you have something to "hide" if you do not. But consider:

It is not unreasonable to suppose that only those who trusted both might be comfortable with intrusive government surveillance. 

Maybe we should trust the government more, or at least trust law enforcement. But many do not.

There is also the issue of self-censorship; see the IETF document below. When subjected to continued surveillance, even if not very intensive, people often have a tendency to speak and act with greater restraint. One might be less likely to attend a protest march, for example, or to criticize the government in emails. Messages and emails might be less likely to express sentiments that disagree with the authorities. There is an incentive to conform. For this to happen, one must, of course, be aware of the surveillance.

You can be very confident you have nothing to hide, nothing to worry about, but still be aware that getting on the No-Fly List would be personally catastrophic.

What is Privacy For?

Closely related to the "nothing to hide" approach is the question of just what privacy is for, socially. Most societies do have strong norms about privacy: some conversations are hushed, for example, when a third party approaches. What is all that about? Is privacy an important social element? Why should we be concerned about what information we share with others? One possibility is that, without privacy, we cannot define our more intimate relationships in terms of sharing additional personal information, because everyone knows everything. Or might the roots of privacy lie in keeping food resources secret?

Sometimes privacy is about, as Judge Richard Posner wrote, the right of a person "to conceal discreditable facts about himself". But privacy clearly goes beyond simply trying to conceal ones past misbehavior.

We do keep passwords private; is that a special case, or is that really a part of privacy? It is one thing to keep our bank password private to prevent others from taking our money, but should we object to others simply knowing how we spend our money?

We also live in spaces with walls, and have curtains on our windows. It is sometimes suggested that sexual privacy is related to relationships that may lack social approval -- and so must be kept private. However, even modest hints of a sexual relationship (eg public displays of affection) often make observers quite uncomfortable.

Sometimes people hold opinions that are contrary to the opinions of those in power, and so keeping those opinions private may be a matter of personal safety.

Privacy is often associated with autonomy and independence; most people value the latter two quite strongly. To lose privacy is also to lose social position; to put it another way, privacy is correlated with social rank. And in essentially all societies, people are very concerned about their social rank. In modern terms, loss of privacy due to government surveillance is indeed associated with feelings of helplessness and powerlessness (though the NSA tried to keep its phone-metadata program very secret). But it is also true that people use privacy to conceal past misbehaviors, and to that extent the purpose of privacy is to manipulate ones social rank so it is higher than it would be otherwise.

Perhaps you have some minor things to hide. Traditionally, that was the justification for the Fourth Amendment. But how does that change in a world with mass terrorist attacks? Some in the NSA have argued that as soon as there is another attack, everyone will be clamoring for more surveillance.

How much surveillance do we need? How much do we want?

On 4e p 50 / 5e p 54, Baase quotes Edward J Bloustein as saying that a person who is deprived of privacy is "deprived of his individuality and human dignity". Dignity? maybe. But what about individuality? Is there some truth here? Or is this overblown?

On 4e p 62 / 5e p 78, Baase quotes Supreme Court Justice William O. Douglas as saying, in 1968,

In a sense a person is defined by the checks he writes. By examining them agents get to know his doctors, lawyers, creditors, political allies, social connections, religious affiliation, educational interests, the papers and magazines he reads, and so on ad infinitum.

Nowadays we would add credit-card records. Is Douglas's position true?

The NSA and the Snowden Leaks

In the aftermath of the September 11, 2001 attack on the World Trade Center, Congress passed the USA Patriot Act (or Usap At Riot Act, as Richard Stallman likes to call it). Title II of this act greatly expanded the powers of federal agencies to conduct surveillance on suspected terrorists.

Congress created the Foreign Intelligence Surveilllance Courts (or FISA Courts) with the Foreign Intelligence Surveillance Act of 1978. These courts gained additional authority with the Patriot Act. The FISA courts were charged with issuing any necessary warrants for surveillance under the Patriot Act.

The NSA eventually began collecting all of the following:
  1. Telephone records of essentially every call placed in the US
  2. contents of emails, Facebook messages, SMS messages and other text-based communications
  3. raw packet data from direct taps into central Internet routers

The NSA claimed that all this was authorized by §215 of the Patriot Act, which allows collection of a wide range of records for investigations involving international terrorism. The pre-9/11 §215 allowed for collection of "business records"; this was amended to allow collection of "any tangible thing". The NSA interpreted this to allow collection of data on US nationals as long as the investigation involved someone who was not a US national. Here is the text of the relevant portion of the act:

(a)(1) The Director of the Federal Bureau of Investigation or a designee of the Director (whose rank shall be no lower than Assistant Special Agent in Charge) may make an application for an order requiring the production of any tangible things (including books, records, papers, documents, and other items) for an investigation to protect against international terrorism or clandestine intelligence activities, provided that such investigation of a United States person is not conducted solely upon the basis of activities protected by the first amendment to the Constitution. [pld: does this provision mean anything?]

(2) .. (b) Each application under this section--

 (1) shall be made to--
  (A) a judge of the court established by section 103(a); or [pld: this is the FISA court]
  (B) a United States Magistrate Judge under chapter 43 of title 28, United States Code, ...
  (2) shall specify that the records concerned are sought for an authorized investigation conducted in accordance with subsection (a)(2) to protect against international terrorism or clandestine intelligence activities.
  (d) No person shall disclose to any other person (other than those persons necessary to produce the tangible things under this section) that the Federal Bureau of Investigation has sought or obtained tangible things under this section.

That the FISA courts are authorized to hear these cases is explicit in (2)(1)(A). That is, the law clearly provides for the FISA courts to authorize release of records of US nationals (the original jurisdiction of the FISA courts was limited to non-US-nationals). The law also makes clear that records can be released as part of any investigation; the person whose records are released does not have to be a subject of that investigation. That is, your records can be released as part of an investigation of someone else.

The last clause quoted here, (d) mandates that communications providers can not reveal to the public or their customers anything about this surveillance activity. These "gag orders" are unpopular with providers. They undermine confidence in the US software-services industry. The reason nobody had any idea about the extent of NSA domestic surveillance before Snowden was that these gag orders prevented talking about it. See below.

In May 2013, Edward Snowden began releasing internal, classified information about the National Security Agency's domestic-spying program. The information was published starting in June 2013 by the Washington Post and the British newspaper The Guardian. While there are lots of spying events documented by Snowden, the two primary ones are the sweep of telephone records and the PRISM program involving the content of emails.

In May 2006, the FISC issued its first order (a mass subpoena) requiring telephone providers to turn over all telephone records to the NSA, as part of the PRISM program. These records include:
The content of the call is not saved. The records above are the normal business records of the providers. The government has long considered "normal business records" to be fair game, although others have objected to this interpretation. The Supreme Court ruled in Smith v Maryland that the police do not need a warrant to gather the called-number information for phone calls. But is your cellphone's location a "normal business record"?

Note that there is no claim by the government that any particular phone number might be associated with illegal activity.

The original order allowed for the collection of the data, but any use had to be approved by the FISC. In 2009 the FISC discovered that the NSA had not been complying with this portion of the requirement. Exactly what is the status of the regulations on the use of this data is not clear.

The PRISM program also involved the collection of contents of email and other text-based messages (and possibly some Skype calls). This data came from providers (eg Gmail, Yahoo and Microsoft). The third leg of the program included data obtained through direct taps into key Internet routers. This information was supposedly collected on a per-name (that is, individual) basis, but emails were included of those who were on the third "hop" away from a suspect (someone who corresponded with a suspect is on the second hop). So communications between US nationals were definitely included.

Supposedly no warrant is needed to monitor communications of either non-nationals or of US nationals traveling outside the US. However, the FISA court generally signed off on the subpoenas involved. Mass surveillance is impractical if "probable cause" must be established for every individual involved, eg for a warrant.

Ten years after the Snowden revelations, the IETF published an Internet Draft,, discussing the consequences of Snowden, and some subsequent developments.

One theory is that warrants are not easy to get, and the relatively lopsided success rate (over 99%) is due to careful preparations by the police for each and every one. There is in fact some evidence that FISA court warrant applications often receive a reasonable degree of care.

Still, there is never anyone involved whose role is to speak against the warrant.

In December 2019 Justice Department's Inspector General released a report about FISA warrants, in particular related to the investigation of Carter Page, a Trump associate. The New York Times published an article:

Generally, FBI agents did not present evidence contrary to the theory that Page was in cahoots with the Russians.

Snowden claimed he tried to bring his legal issues with mass surveillance to the attention of his superiors at the NSA. The NSA denied this. A 2016 article suggested that the NSA was not being truthful: However, neither did Snowden present detailed descriptions of his attempts at contacting his superiors.


The US is not the only country to engage in broad surveillance. Denmark offers extensive social-welfare benefits, spending a remarkable 26% of budget on this. Some Danes have become very concerned that these benefits are not going to those who are not entitled to them. To this end, the Danish government has created a group within the Public Benefits Administration to root out cheaters, using surveillance techniques. See

The anti-fraud group cross-checks multiple government databases, looking, for example, for benefit recipients who are employed, or who travel frequently, or who own cars. Such matching has become routine in the US as well. The group has also considered checking electric and water bills to see if recipients were actually living at the addresses they listed. The system does track relatives in other countries, both in the EU and outside.

There is particular concern that someone receiving benefits might in fact be living with a partner who does not qualify. To detect this, nearest-tower cellphone location data has been used to determine where someone is sleeping. Child registries are also used to identify the other parent of any of a recipient's children, to see if the benefit recipient is living with that other parent. Credit-card records are checked, to see, for example, if there are regular gas and food purchases near ones claimed home address. And there are in-person surprise visits, though those are expensive. There are claims that recipient's social-media histories have been searched for this purpose.

See also (in Danish, but Google Translate does a fair job).

The Danish system is supposed to allow a hearing for everyone accused. Apparently these hearings impose penalties on only 8% of those flagged, suggesting that 92% had done nothing wrong.

A similar system in the Netherlands falsely flagged thousands of families for welfare fraud in 2021. It eventually turned out that the algorithm used national origin inappropriately; the people flagged were mostly recent immigrants.


If your interaction with Facebook or gmail was via https, that is, via an encrypted web connection, then the NSA would have to decrypt anything it obtained through router taps. Decryption of much https traffic is not terribly difficult, but it is time-consuming, and the NSA probably cannot afford to decrypt all of it. Obtaining message information from the providers -- such as Facebook and Google -- avoids that.

You can encrypt your email on an end-to-end basis, but that is not exactly trivial. The standard open-source public-key encryption package is probably GPG (Gnu Privacy Guard). There is a plugin for the Thunderbird email reader, known as enigmail, that provides email support for GPG. That is, email messages to and from selected recipients are automatically encrypted and decrypted.

Catch #1: You have to resolve the public-key-trust issue. Suppose Alice wants to email Bob, with whom she has no pre-existing relationship. Then Alice needs Bob's public key. She can just trust that the key is the one on Bob's website, but what if the NSA redirects Alice to a fake copy of Bob's site, with a fake public key? Alice then sends the email encrypted with the NSA's public key. The NSA decrypts it, saves it, and re-encrypts it with Bob's real public key and delivers it to Bob. Bob is none the wiser.

This is known as the "man-in-the-middle" attack.

The traditional assumption here is that you get other people's public keys from people you trust. This can be tricky.

The Signal encrypted-text-message system has a reasonably convenient approach to this problem. If Alice is worried, she can call Bob (the idea is that she would recognize Bob's voice) and the two can exchange key "fingerprints" by voice.

Catch #2: How many other people will set up encryption? Until there is a large number, Alice's email stands out by dint of being encrypted. The NSA can devote intense resources to breaking the encryption. And Alice is now on the Watch List.

Also, you can only use encryption with other people who have set it up. Most of your email is thus likely to remain plaintext.

Signal has tried hard to make encryption universal. Their biggest success was probably in convincing Whatsapp to use their TextSecure protocol.

Catch #3: Where do Alice and Bob keep their keys? If they are on their respective computers permanently, then they are vulnerable. If they are only entered when necessary, then the act of typing the key is a weak point. If Alice and Bob want to get each others' email on the go, and try to use encryption on their smartphones, that becomes a weak point.

Parallel Construction

On the one hand, national security is an important goal. But what about the following two-step argument:
There have been repeated claims that the Special Operations Division of the DEA has been beneficiary of some NSA data, and has been using it in narcotics arrests. DEA agents, according to this theory, have been trained in the art of parallel construction -- coming up with an alternative explanation for why someone was arrested, that avoids disclosure of the NSA data. While to a point this is legitimate, ultimately the defendant's right to a fair trial depends on obtaining all information about how a case was investigated.

More disturbingly, use of personal data obtained without a warrant is often forbidden at trials. If the NSA/DEA subterfuge here acually occurred, then it intentionally bypasses that. The NSA has also shared information with other federal agencies, including the IRS.

The effect of all this is would be to allow the use of NSA-collected data in ordinary criminal prosecutions.

See the article at However, the facts here are difficult to verify, so we don't really know the extent to which parallel construction was used. The DEA apparently does benefit from parallel construction, often to protect informants. The part that is not clear is the extent to which the NSA dragnet has been the source.

A more recent, better-documented example [2014] is described in A memo regarding the use of the Stingray cellphone tracker was sent to the Oklahoma City police department by FBI agent James Finch. It reads, in part:

Information obtained through the use of the equipment is FOR LEAD PURPOSES ONLY, and may not be used as primary evidence in any affidavits, hearings or trials. This equipment provides general location information about a cellular device, and your agency understands it is required to use additional and independent investigative means and methods, such as historical cellular analysis, that would be admissible at trial to corroborate information concerning the location of the target obtained through the use of this equipment.

The problem here is that it is illegal for police to withhold evidence, but it is hard to read this paragraph as not advocating just that! If historical cellular data is adequate, why do the police need a Stingray? The real problem with historical cellular data is that the police use the Stingray to identify the suspect; finding the applicable historical data is like looking for a needle in a haystack.

In the last century, the federal government discouraged encryption with the stated goal of being able to investigate the following groups:

What happens when the third group above includes recreational (or medical) marijuana users? For that matter, what if the first group is taken to include anyone who expresses an interest in a "subversive" organization, such as Occupy Wall Street?

In 2016, the FBI petitioned Congress for access to a basic form of browser records: a list of what web IP addresses you connected to. The FBI has tried to argue that this was left out "by mistake" from a much earlier version of the law, but that law explicitly listed only telephone records. See

Is a telephone call record really like a website connection? The phone company uses the former for billing (or used to); no ISP uses your website connections for billing. In this sense they are not business records.

In 2008, Yahoo attempted to fight the PRISM-based FISA Court order to turn over a large volume of emails. The case made it to the appellate level -- the US Foreign Intelligence Surveillance Court of Review. The partially redacted decision -- at -- is chilling. First off, the appeals court makes several references to the trial court decision, but the trial court decision is secret. The trial-court decision is referred to as "Sealed Case".

Yahoo raised two issues: that a warrant was always needed, even for foreign nationals, and that the requests for information were "unreasonable".

The first point -- whether the fourth amendment applies to foreigners -- is a serious issue, but the court dismissed it without considering precedent. As for the second point, the court basically agreed that there should be no clear line "between foreign intelligence purposes and criminal investigation purposes". Of course, some might argue that this should raise the bar for whether a search was "reasonable", but the FISC ruled that as long as the stated purpose was foreign-intelligence gathering, then subpoenas were ok. The FISC turned the Fourth Amendment on its head by then arguing (p 17) that warrants were "unreasonable":

We add, moreover, that there is a high degree of  probability that requiring a warrant would hinder the government's  ability to collect time-sensitive information and, thus, would impede the vital national security interests that are at stake.

Finally, the court decided that whether a search was "reasonable" must depend on its importance. If national security is at stake, essentially all searches (according to the opinion) become reasonable.

At one point (page 12) the decision states, in case the reader is confused, "This makes perfect sense".

Spying and harm

Does the NSA spying on Americans actually cause any harm to ordinary Americans? Is it true that if we have nothing to hide, then we have nothing to fear?

The government has long kept tabs on those who participate in protest movements. So what?

Is there a "chilling effect"? If so, is it strong enough to matter?

According to a Congressional investigation committee, "Martin Luther King, Jr. was the target of an intensive campaign by the Federal Bureau of Investigation to 'neutralize' him as an effective civil rights leader." What could the FBI actually do to MLK? They tried exposing him as a communist, but failed as MLK had no ties to communism.

In November 1964 the FBI sent King an anonymous letter, here, in which the letter writer threatens to expose King as a fraud (possibly for adultery) and suggests that the only way out is for him to commit suicide.

Alternatively, perhaps the government might have tried blackmailing King.

Is this concern of any large-scale significance?

How is this related to apparent NSA use of sexual information to discredit what it calls "radicalizers"?

The National Security Agency has been gathering records of online sexual activity and evidence of visits to pornographic websites as part of a proposed plan to harm the reputations of those whom the agency believes are radicalizing others through incendiary speeches, according to a top-secret NSA document. The document, provided by NSA whistleblower Edward Snowden, identifies six targets, all Muslims, as 'exemplars' of how 'personal vulnerabilities' can be learned through electronic surveillance, and then exploited to undermine a target's credibility, reputation and authority.

Is this the basis for another argument that we are entitled to privacy even if we have "nothing to hide"? On the other hand, this case is 60 years ago, and King did not commit suicide.

Do we agree to this?

James Clapper, director of the NSA, says "We Should've Come Clean About Phone Surveillance":

I probably shouldn't say this, but I will... Had we been transparent about this from the outset right after 9/11 -- which is the genesis of the 215 [Section of the Patriot Act -- pld] program -- and said both to the American people and to their elected representatives, we need to cover this gap, we need to make sure this never happens to us again, so here is what we are going to set up, here is how it's going to work, and why we have to do it, and here are the safeguards ... We wouldn't have had the problem we had.... If the program had been publicly introduced in the wake of the 9/11 attacks, most Americans would probably have supported it.

Never mind that in June 2013 when the phone surveillance first came to light he was quite upset that the secrecy of the program was lost. Now that our enemies knew about it, he said then, they would find other ways to communicate.

In the post-9/11 context, do you support at least the basic framework of the NSA surveillance? Do you think a majority of Americans do? There may have been some excesses (such as hacking and "parallel construction"), but ignore those for the moment.

Smart cities

Some cities are trying to become "smart cities", leveraging information technology to become better places to live. Here's an article about Toronto: Sensors collect information about car usage, bike usage, pedestrian usage, and trash-can usage.

This might lead to better traffic flow, but at what privacy cost? Nobody knows.

ACLU v Clapper

On May 7, 2015, the Second Circuit released their decision in ACLU v Clapper, in which they found that Section 215 of the Patriot Act does not allow bulk phone-metadata collection. Implementation of the ruling was stayed, however, pending appeal to the Supreme Court.

The decision did not address whether such collection violates the Fourth Amendment; the claim was simply that the existing Section 215 did not allow for the data collection that was being done.

The opening of the argument raised issues of domestic FBI surveillance during the 1970's, which was eventually significantly curtailed.

The court also pointed out

A call to a single-purpose telephone number such as a "hotline" might reveal that an individual is: a victim of domestic violence or rape; a veteran; suffering from an addiction of one type or another; contemplating suicide; or reporting a crime. Metadata can reveal civil, political, or religious affiliations; they can also reveal an individual's social status, or whether and when he or she is involved in intimate relationships.

A large part of the case hinged on whether the ACLU, together with a set of telephone subscribers, had in fact standing to sue. The Second Circuit held that they did, because the government had collected their phone records. Actual use of the records did not have to be shown, let alone actual harm.

On June 1, 2015, section 215 of the Patriot Act expired, along with a few other provisions.

The next day Congress passed (and the president signed) the so-called Freedom Act, which granted a 6-month extension to the NSA's phone-metadata-collection program. After that time, the data-collection program apparently came to an end.

On June 29, 2015 the FISA Court of Appeals ruled that the Freedom Act had implicitly authorized the continuation of the NSA's metadata-collection program, at least for 6 months, and thus "reversed" the Second Circuit. The reversal of the Second Circuit decision raises a decidedly awkward question of jurisdiction, but the FISA court has a point, and the Second Circuit had stayed their own order pending appeal.

But because of the ending of the bulk-data-collection program, the case was not appealed to the Supreme Court.

Microsoft vs US

(This case is also known as the "Microsoft Ireland case", and is not to be confused with the antitrust litigation US v Microsoft.)

In December 2013, Microsoft received a search warrant from the US Department of Justice for the email of a drug-trafficking suspect. Microsoft refused, on the legal theory that the data was stored at a data center in Ireland, and that therefore Irish laws should apply. The DoJ, instead of obtaining approval of Irish authorities, decided that Microsoft had to turn over the data because it was a US company, no matter where the data was located.

One problem with the government's legal theory is that it may be illegal in the remote jurisdiction to turn over documents without a warrant.

See and also

Microsoft appealed to the Second Circuit. In July 2016 a three-judge panel ruled unanimously in Microsoft's favor: the US must obtain a warrant in Ireland, under the existing mutual-legal-assistance treaty. The US asked for an en banc rehearing. This motion was denied in January 2017; the eight circuit judges ruling on the motion were split four to four, and so the three-judge panel decision holds.

The case was the highlight of the 2017 Supreme Court season, but Congress passed the Cloud Act in March 2018. The Supreme Court dropped the case the following month, as moot.

The Cloud Act means that, in general, a US warrant for information must be honored by a US provider no matter where in the world the data is stored. However, the provider can object if the provider believes that turning over the information would violate the privacy laws of the hosting country. It also allows for the negotiation of international agreements for the fast turnover of such information; these agreements are not treaties and so are much easier to negotiate.

A big concern for Microsoft -- and other US companies -- had been that if the DoJ had prevailed, then foreign companies would likely be increasingly reluctant to trust US-based cloud providers -- even when the cloud storage is physically located outside the US. This case, therefore, was central to Microsoft's business interests.

In a related case, a Microsoft employee was charged in Brazil with failure to turn over Skype records. In Brazil, turning over the records was required, but at the same time in the US turning over the Brazilian records was (and still is) forbidden.

Supreme Court cases on privacy
-- Baase 4e pp 63ff / 5e p 77

1928: Olmstead v United States

The Supreme Court ruled that federal agents did not need a warrant when they tapped Roy Olmstead's phone.

In 1934 Congress passed the Communications Act, which created the FCC and which also banned [in Section 605] telephone wiretaps without a warrant. Yet the law was awkwardly worded, and wiretaps by private investigators continued.

1967: Katz v United States
The 4th amendment does too apply to wiretaps! Privacy may still exist in a public area.

What a person knowingly exposes to the public, even in his own home or office, is not a subject of Fourth Amendment protection. See Lewis v. United States; United States v. Lee. But what he seeks to preserve as private, even in an area accessible to the public, may be constitutionally protected.

Katz was using a pay phone; the FBI had a microphone just outside the phone booth. To the appellate court, the fact that the microphone did not intrude into the phone booth was significant in finding for the FBI, but the Supreme Court reversed.

Under Katz, the doctrine of "reasonable expectation of privacy" (REoP) replaced the doctrine of "physical intrusion".

The problem with the REoP doctrine: as technology marches on, isn't our reasonable expectation diminished? And does this then give the government more license to spy?

Note the first part of the quotation above: if you expose something to the "public", it is not private. This was later formalized in the Miller decision, next, despite the following also from the Katz decision:

Indeed, we have expressly held that the Fourth Amendment governs not only the seizure of tangible items, but extends as well to the recording of oral statements, overheard without any "technical trespass under . . . local property law." Silverman v. United States, 365 U. S. 505, 365 U. S. 511. Once this much is acknowledged, and once it is recognized that the Fourth Amendment protects people -- and not simply "areas" -- against unreasonable searches and seizures, it becomes clear that the reach of that Amendment cannot turn upon the presence or absence of a physical intrusion into any given enclosure.

This second quote strongly suggests that your "papers" do not have to be physical, or under your direct control, to be covered by the Fourth Amendment.

Between Olmstead and Katz, there had been a gradual recognition of increasing scope of the Fourth Amendment, hence the thought on the part of the Katz defense team that this was worth pursuing.

1976: US v Miller 425 US 435

(There are at least three major Supreme Court cases involving someone named "Miller").

Miller's incriminating bank records were subpoenaed. Miller tried to argue that a warrant was needed. In this he lost.

The Supreme Court ruled that information we share with others (eg our bank) is NOT private. The government can ask the bank, and get this information, without a warrant. (However, the bank could in those days refuse.)

The Fourth Amendment does not prohibit the obtaining of information revealed to a third party and conveyed by [that third party] to Government authorities

Justice William O Douglas was quoted earlier as saying a person could be "defined by the checks he writes". Douglas might not have agreed with the Miller decision, but he died in 1975.

Third-party doctrine

The Miller decision created what is now known as the third-party doctrine: all "business records" about us are fair game for an ordinary subpoena. On the one hand, this is a straightforward extension of the idea in Katz that what you expose to the public is not private (though there is room to debate just what is "public"). On the other hand, though, Miller had tried to use the second Katz quote above in his defense, that papers don't have to be physically under ones control, and lost.

The decision quoted from an earlier ruling

Banks are . . . not . . . neutrals in transactions involving negotiable instruments, but parties to the instruments with a substantial stake in their continued availability and acceptance.

This suggests that the transaction theory (later) of privacy is involved: both parties have significant interests in the records. What about "business records" that are largely irrelevant to the operation of the business? Email providers have zero involvement in the content of the email (except for gmail?), and cellular providers have no interest in your nearest-tower location after you have left that particular cell. Unfortunately, the Supreme Court has never really addressed this aspect of the Third-party Doctrine, or for that matter even spelled out a constitutional justification for it.

Another thread in the third-party doctrine comes from informants wearing a wire. The Supreme Court ruled in On Lee v US that no warrant was needed for that, whereas a warrant would be needed if the informant were not present and the wire were simply an eavesdropping device. The argument here is that the wire simply records what the informant -- the third party -- has heard directly.

For a good history of the third-party doctrine by Orin Kerr, see

We will continue with this here in the context of email.

1979: Smith v Maryland

Reduction of REoP by the police is not SUPPOSED to diminish our 4th-amendment rights. However, in that case the Supreme Court ruled that "pen registers" to record who you were calling did NOT violate the 4th amendment.

Patricia McDonough had her purse stolen. She remembered the assailant's car. Soon after, she began receiving crank calls, and recognized the car driving down her street. A police officer saw the car, noted its license plate, and discovered the car was registered to Michael Smith. A pen register was placed on Smith's home line; this revealed calls to McDonough. Based on those calls, the police got a warrant, and at that point found further evidence. Smith argued, through his lawyers, that the pen register was a warrantless search and that all the later evidence should be thrown out. He lost.

Application of the Fourth Amendment depends on whether the person invoking its protection can claim a "legitimate expectation of privacy" that has been invaded by government action. This inquiry normally embraces two questions: first, whether the individual has exhibited an actual (subjective) expectation of privacy; and second, whether his expectation is one that society is prepared to recognize as "reasonable."

First, we doubt that people in general entertain any actual expectation of privacy in the numbers they dial. All telephone users realize that they must "convey" phone numbers to the telephone company, since it is through telephone company switching equipment that their calls are completed. All subscribers realize, moreover, that the phone company has facilities for making permanent records of the numbers they dial....

If you want to keep a number private, don't call it!

Note the crucial issue that the defendant voluntarily shared the number with the phone company! Of course, if you want to use a phone, you have no choice.

Justices Stewart & Brennan dissented

The telephone conversation itself must be electronically transmitted by telephone company equipment, and may be recorded or overheard by the use of other company equipment. Yet we have squarely held that the user of even a public telephone is entitled "to assume that the words he utters into the mouthpiece will not be broadcast to the world." Katz v. United States

What do you think of this distinction? Is there a difference between sharing your phone number with the phone company and sharing your actual conversation with them? Is the phone number a "business record" of continued relevance? How does the phone number (which at the time of the case would have been used for billing) differ from a cell-tower location? After all, even today cell-tower locations are used to determine whether you are roaming, and thus affect your bill.

Do you think the Supreme Court might have answered differently if they had envisioned NSA-type "pen registers" on essentially everyone in the United States? Note that Smith was an active suspect; the police probably could have obtained a warrant based on McDonough's tying of Smith's car to her robbery.

The Smith case represents a further extension of the third-party doctrine to calling records.

2001: Kyllo v United States

Thermal imaging of your house IS a 4th-amendment search! This is a very important case in terms of how evolution in technology affects what is a REoP

Held: Where, as here, the Government uses a device that is not in general public use, to explore details of a private home that would previously have been unknowable without physical intrusion, the surveillance is a Fourth Amendment 'search', and is presumptively unreasonable without a warrant.

How long into the future will this hold? Could it be that part of the issue was that the general public was not very aware of the possibility of thermal imaging? If thermal imaging were to come into not only general public awareness but also general public use (eg by equipping cellphones with IR cameras), would the situation change?

In 2016, Caterpillar (yes, the maker of the D11 Bulldozer) has now introduced a phone with an infrared camera:

In 1990 the Supreme Court let stand a lower-court decision that eavesdropping on someone else's phone call made on an old-fashioned cordless phone (remember those?) was not an invasion of privacy because no one had a "reasonable expectation of privacy" with these devices. Many users did know that it was easy to listen in to someone else's call simply by playing with the channel button. See

A related issue came up in the context of John and Alice Martin's 1996 taping of an embarrassing conversation involving then-Representative Newt Gingrich, who was engaged in a frank discussion of some ethics lapses. The Martins used a police scanner to listen in to Rep. Gingrich's "cell" phone; the phone was likely a first-generation analog (or "AMPS") model that was almost as easy to eavesdrop on as a cordless phone though this did require special equipment. The Martins were eventually fined $1,000.

Jonas Diener

(This was not a Supreme Court case.) Jonas Diener of Virginia was using eight times the normal amount of electricity at his home. Based on that, police obtained a warrant, believing he was running a marijuana "grow house". They did find some marijuana, but it was unrelated to the electric usage. Diener was not growing marijuana. The electricity use was due to a large-scale bitcoin-mining server Diener had set up.

Diener received a six-month suspended sentence. Initially the police seized his computer hardware  and his bitcoins.

In general, once a search warrant has been executed, it is still possible to challenge the search by making a motion to suppress evidence obtained from the search. (Sometimes this is called a motion to quash, though apparently that is really supposed to apply only to warrants that have been issued but not executed.) Diener could have argued that excessive electrical usage is not probable cause for a drug search -- his own bitcoin-mining operation would have been Exhibit A here -- and there is a good chance he would have prevailed. 

However, justice like that is expensive. It appears Diener settled for the suspended sentence rather than fighting the legal principles. The fact that the government offered a completely suspended sentence suggests that they were worried at least a little about losing the case.

2012: United States v Antoine Jones

Jones was an alleged cocaine dealer in the Washington, DC area. Police attached a GPS tracker to his car while it was parked in the driveway. By following him over a 30-day period, the police were able to build a strong case against him. But Jones argued that such tracking was unreasonable warrantless search, despite a 1983 Supreme Court ruling that allowed wireless tracking for single trips. The Department of Justice argued that no one has a REoP regarding his or her movements on public streets. The DoJ also pointed to the 1983 US v Knotts case in which police had the manufacturer attach a radio beeper to a drum of chloroform. When Knotts purchased the drum, police used the beeper to track him to his cabin in the woods.

In August 2010, the DC Court of Appeals agreed with Jones, and overturned his conviction. (This decision was known as US v Maynard.)

The ninth circuit and the seventh circuit (including Illinois) had ruled otherwise, however.

The Supreme Court ruled unanimously in January 2012 that "the Government's attachment of the GPS device to the vehicle, and its use of that device to monitor the vehicle's movements, constitutes a search under the Fourth Amendment." As such, a warrant would be required.

However, by 5-4 the court also ruled that the issue here was the government's trespass onto private property to install the GPS tracker. That is, the court did not rule broadly (by explicit choice!) on the question of whether sustained GPS tracking itself violated a person's reasonable expectation of privacy. Justice Scalia wrote the majority opinion, arguing that rules against government trespass should coexist with the REoP approach, and that this particular case could be decided on trespassing grounds without the need to consider REoP (which others on the court agreed was a problematic standard). Note that the trespass ruling makes the decision consistent with Knotts.

Jones was tried again in January 2013; in that trial, the government used nearest-tower location data instead of GPS data. That trial ended in a hung jury. The government prepared for yet another trial, but Jones finally accepted a plea bargain of 15 years with credit for time served.

In US v Katzin, 2013, the Third Circuit ruled that the police must obtain a warrant simply to monitor GPS trackers. In this case, the device was installed before the US v Jones decision, but the police continued to monitor the device afterwards. The Third Circuit ruling expressly states that a warrant is required both to install a GPS tracker and to monitor it.

The Mosaic Theory

In the DC Circuit version of the Antoine Jones case (US v Maynard), the court developed what they called the "mosaic theory": that one individual record might not require a warrant, but that continued use of such data could be a different story.

[W]e hold the whole of a person's movements over the course of a month is not actually exposed to the public because the likelihood a stranger would observe all those movements is not just remote, it is essentially nil. It is one thing for a passerby to observe or even to follow someone during a single journey as he goes to the market or returns home from work. It is another thing entirely for that stranger to pick up the scent again the next day and the day after that, week in and week out, dogging his prey until he has identified all the places, people, amusements, and chores that make up that person's hitherto private routine.

... When it comes to privacy, however, precedent suggests that the whole may be more revealing than the parts.

What do you think of this? In many ways, this is the heart of the NSA-surveillance issue: that the NSA took rules allowing isolated surveillance, and applied them universally.

Here is the Volokh Conspiracy's take on this (by Orin Kerr).

The FBI and cellphone location records

nearest-tower (cell-handoff) records v GPS records

Supposedly the Justice Department gets warrants for GPS data (nearest few feet), but usually does not for nearest-tower data (which positions you to within a few miles at worst, a few hundred feet at best).

Another distinction is between realtime data (where you are now) and "historical" data (where you were over the past month).

The federal government has tried to claim that nearest-tower data simply amounted to "routine business records". Are they?

Note that the Jones opinion does not apply here as the police do not trespass when they acquire GPS phone records.

Melvin Skinner

In August 2012 the Sixth Circuit ruled that police could use GPS tracking of Melvin Skinner's phone without a warrant. Skinner was arrested with 1,100 pounds of marijuana in 2006; he was tracked via his prepaid cellphone.

The decision stated that the Antoine Jones case was specifically about police intrusion onto Jones' property to attach a tracker, and not about GPS tracking per se.

However, in Melvin Skinner's case, the police (with the cooperation of the cellphone provider, under court order but not a warrant) turned on the GPS feature of the phone remotely. Skinner was not tracked through his use of the phone to make calls, or because he enabled GPS himself.

Judge John Rogers wrote,

When criminals use modern technological devices to  carry out criminal acts and to reduce the possibility of detection, they can hardly  complain when the police take advantage of the inherent characteristics of those very devices to catch them.

In short, Skinner did not  have a reasonable expectation of privacy in the data emanating from his cell phone that  showed its location.


There is no Fourth Amendment violation because Skinner did not have a reasonable expectation of privacy in the data given off by his voluntarily procured pay-as-you-go cell phone.  If a tool used to transport contraband gives off a signal that can  be tracked for location, certainly the police can track the signal.  The law cannot be that a criminal is entitled to rely on the expected untrackability of his tools. [1] Otherwise, dogs  could not be used to track a fugitive if the fugitive did not know that the dog hounds had  his scent.  A getaway car could not be identified and followed based on the license plate  number if the driver reasonably thought he had gotten away unseen. The recent nature  of cell phone location technology does not change this. If it did, then technology would  help criminals but not the police.  It follows that Skinner had no expectation of privacy  in the context of this case, just as the driver of a getaway car has no expectation of  privacy in the particular combination of colors of the car's paint.

The footnote [1] clarifies:

We do not mean to suggest that there was no reasonable expectation of privacy  because Skinner's phone was used in the commission of a crime, or that the cell phone was illegally possessed. On the contrary, an innocent actor would similarly lack a reasonable expectation of privacy in the inherent external locatability of a tool that he or she bought.

On the other side of the question, what is so hard about getting a warrant? In Skinner's case, it appears that the police already had plenty of "probable cause".

A significant problem with the Skinner decision is that, while the police discovery of Skinner's location clearly depended on GPS data, most of the decision's argument makes sense only for the case of nearest-tower data. Note, for example, the quote above where it says "Skinner did not have a reasonable expectation of privacy in the data given off by his voluntarily procured pay-as-you-go cell phone"; it is nearest-tower data that is given off "volunarily". There are suggestions that the Sixth Circuit simply confused the two.

Another issue with the decision was that the Sixth Circuit declared there was no "trespass" and so the Jones precedent did not apply. But remotely accessing a phone, enabling the GPS service, and sending the GPS coordinates back is a very "active" step; such steps have been ruled as "trespass" in the past, in hacking cases. The Sixth Circuit did not consider whether this action by the provider might be a form of trespass.

The Skinner decision does represent something of a "circuit split": other appeals courts have since ruled against this kind of GPS tracking. In 2015, in US v Graham, the Fourth Circuit ruled that police must obtain a warrant even to make use of real-time nearest-tower data.

However, the Supreme Court has not yet resolved the issue.

Riley and Wurie

In 2014, the Supreme Court ruled in the combined cases Riley v California and US v Wurie that the police do need a search warrant to search a cellphone in the possession of someone who is arrested. This was a unanimous decision, written by Justice Roberts, with Justice Alito writing a concurring opinion. Roberts wrote

Modern cell phones are not just another technological convenience. With all they contain and all they may reveal, they hold for many Americans "the privacies of life". The fact that technology now allows an individual to carry such information in his hand does not make the information any less worthy of the protection for which the Founders fought.

In his concurrence, Alito pointed out that this decision now means that police can search any printed records they find on someone arrested, but not any electronic records. Alito thought that complicated distinctions like this should really be made by Congress.

Prior to this case, state courts had struggled with this issue. See below.


In 2018 the Supreme Court ruled, 5-4, that the government needs a warrant to access historical cell-phone records (including nearest-tower records).

The decision was written narrowly, so as not to overturn the third-party doctrine generally. Note that cell-phone records are indeed third-party records. Other third-party records, such as those generated by automatic license-plate readers, remain available without a warrant.

It is possible to read the Carpenter decision as noting that, during the era when the Miller and Smith cases were decided, phones were all landline, and that just because we're all now using cellphones does not mean that the government automatically gets new investigative capabilities.

The decision also did not address real-time nearest-tower records.

As far as I can tell, cellular companies can still voluntarily turn over the data. Though they probably do not, so as to avoid getting in trouble for that.

Melvin Skinner was convicted on the basis of real-time GPS records.

The main workaround by the police at this point appears to be buying historical location information from third parties. Usually this information comes from gps-based or Wi-Fi-based tracking of phones (eg by Google); this tracking does not involve cell-tower records at all. It is also usually more accurate.

Customs and Border Patrol is a big buyer, see

And Fog Data Science is a big provider to police departments (and anyone else who can print up a letterhead with the word "security" in it). The EFF did a study of them:

Apparently Google Maps does not share data with third-party brokers, but Waze does. Even though Google owns Waze.

Sometimes police use "geofence" warrants to Google, to obtain, for example, the location information on anyone (who has their location service enabled) who was within four blocks of the bank within half an hour of the robbery. See #geofence. Google usually responds initially with anonymized data tracks, and then identifies phones only after the police identify some tracks as particularly interesting. Fog's service allows police to track however they want; it also allows police to track locations of phones at, for example, protest events.

See also privacy_others.html#location2, for commercial, non-police use of location data.

Video surveillance

Baase 4e Section 2.2.4 / 5e Section 2.4.1

This is a big issue in Chicago, where there are both "obvious" and "hidden" cameras.

2001 Super Bowl: Tampa police used facial-recognition software on all 100,000 fans. The software did not work terribly well.

London: heavy camera use to:

London in 2005:
What about the rate of false positives? This is diminishing, slowly but surely.

Should the London cameras be used to track lesser crimes, such as pickpocketing? Supposedly the Chicago street cameras have been quite effective in handling minor crimes.

The police can track you outside your home by any or all of the following:
Is it true that you have no expectation of privacy outside your home?

There is a longstanding notion that one can be anonymous in a large crowd; is there a legal basis for this?

After all, if you are seen on the street by someone who recognizes you, or your car, your privacy is lost. The police can simply follow you, if they are interested in you.

Most people attending protest rallies perhaps assume that the police will not recognize them. What if the police can easily look everyone up? Should there be controls in place?

Facial recognition is starting to get quite accurate, particularly if the government has multiple photos of you. A huge trove of tagged photos, for example, can be found at Facebook. The Clearview application has harvested many such photos, and made them available for use by the police (and some others). See privacy_others.html#clearview.

Consider again the California DOT incident in which the state sent letters to everyone whose car was recorded using a certain stretch of highway. People at the time were upset. Should they have been informed that this was a public road and they simply needed to get over it? Would that be what would happen if the incident occurred today?

Baase 4e p 86 / 5e p 91: case study on federal DB on all US college students. The database would list all courses taken, with grades; it would also include loan and scholarship records.

This is a good example of a fairly common situation: creation of a new database containing confidential information.



Is such a database a good idea?

What if in 2025 a law is passed giving prospective employers access to the data, if the job applicant signs a consent form? What do you think would happen if you refused to sign?

And here's a related issue sometimes called "database-matching": should the government be able to link databases of:

Once upon a time, this was claimed to be an unfair "search". Is it?

Geofence warrants

In the past few years, police have increasingly been relying on the following:

Both of these represent a marked departure from traditional warrants, in that no suspect is explicitly named, and so of course there is no probable cause that the (unnamed) suspect is guilty. Perhaps the most fundamental aspect of the Fourth Amendment is that it requires that the person and place to be searched be described specifically; geofence and search-keyword warrants fail to meet this standard. While both warrants here have value in police work, making them compliant with the Fourth Amendment is not quite yet resolved.

For a summary, see

For a bit more legal detail about geofence warrants see

One issue with geofence warrants is that they sweep up many innocent people, with a special emphasis on joggers, bicyclists and late-night walkers. Another issue is that they only collect information on people who have Google's location-tracking service turned on.

Sometimes people are just questioned, but Jorge Molina was arrested and held for nearly a week. The HLR article states "As a result, Molina dropped out of school, lost his job, car, and reputation, ...". That is a terrible outcome. And the police should have realized immediately that Molina should not have been a suspect.

As for warrants for search history, we will visit the early, local-computer version of this at privacy_others.html#google, but since then a person's search history has moved entirely to Google, away from the local computer. But, like geofence warrants, search-history warrants sweep up everyone (perhaps everyone in the country) who searched for a specific term.

Gag orders have made keeping tabs on police use of search-history warrants difficult. But one such order was uncovered through a court mistake; see The police asked for information about everyone who searched for a woman who later went missing. Google has not released how many accounts were involved.

Geofence warrant requests have been getting broader; Google's ability to push back has been limited. Here is an example involving 1500 users:

Ross William Ulbricht, aka Dread Pirate Roberts, was arrested on October 1, 2013 for his role in running the hidden Silk Road website, generally reached only via onion routing but possibly at, and for his efforts to hire an assassin using the site. The Silk Road website itself, widely used for purchasing illegal drugs, was also shut down.

One of the biggest FBI leads was the interception by the Canadian government of nine fake drivers' licenses ordered by Ulbricht. Another may have been the following post on, about programming and onion routing:

Does the Silk Road have any legitimate uses?

Electronic Communications Privacy Act, 1986

You might think that, because the Supreme court readily recognized that a warrant is needed for phone wiretapping, the same would hold true of email.

This is not, however, how the government saw it.

The difference, for government lawyers, is that email sits around after it is read. At that point it is a "stored document", and maybe even a "business record" of your ISP. It has taken a long time to achieve any form of recognition of a  "reasonable expectation of privacy" for email.

The ECPA was intended to provide protections for "electronic communications". In 1986, Congress would have been aware of email, but it has been suggested that the real target was voicemail. And faxes. Email would have been a distant third.

Part of the intent of the ECPA was to extend the existing restrictions on government telephone wiretaps to other electronic communication. However, the ECPA also applies to private organizations. It has three exceptions that serve to limit its applicability to employer monitoring (§2511(2)(a))

  1. The provider exception (except  that a provider .. shall not utilize service-observing or random monitoring except for ...quality control checks)
  2. The ordinary course of business exception
  3. The consent exception. (c)

Generally, most employer monitoring falls under one of these. Note that the "provider" exception is a specific feature of ECPA; ownership of the hardware does not create a general right of access and in particular ownership of a telephone system does not create a right to eavesdrop.

Phone surveillance in the workplace
Keystroke monitoring
Location monitoring

Do computers empower workers, or shackle them?

While we're on the topic of ECPA, there is:

Title I, covering electronic communications in transit (USC Title18 Chapter 119)
       no interception (§2511):
       evidence exclusion (§2515):

The language on "interception" is below.

(1) Except as otherwise specifically provided in this chapter any person who—

(a) intentionally intercepts, endeavors to intercept, or procures any other person to intercept or endeavor to intercept, any wire, oral, or electronic communication;

shall be punished as provided in subsection (4) or shall be subject to suit as provided in subsection (5).

Title II, the Stored Communications Act. (USC Title 18 Chapter 121)

    email stored 180 days or less: gov't needs a warrant
    more than 180 days: warrant, subpoena, or court order. A warrant is needed if there is no notice to the customer.
        §2703 (a): less than 180 days (b): more than 180 days

Here's that part from §2703(b) preserved for posterity:

A governmental entity may require the disclosure by a provider of electronic communications services of the contents of a wire or electronic communication that has been in electronic storage in an electronic communications system for more than one hundred and eighty days by the means available under subsection (b) of this section.
(b) Contents of Wire or Electronic Communications in a Remote Computing Service.
(1) A governmental entity may require a provider of remote computing service to disclose the contents of any wire or electronic communication ...
(A) without required notice to the subscriber or customer, if the governmental entity obtains a warrant issued using the procedures described in the Federal Rules of Criminal Procedure (or, in the case of a State court, issued using State warrant procedures) by a court of competent jurisdiction; or
(B) with prior notice from the governmental entity to the subscriber or customer if the governmental entity
(i) uses an administrative subpoena authorized by a Federal or State statute or a Federal or State grand jury or trial subpoena; or
(ii) obtains a court order for such disclosure under subsection (d) of this section; except that delayed notice may be given pursuant to section 2705 of this title.

§2703(b) (above) has much less stringent restrictions. Debate continues as to the appropriate category for email messages, but see below under Warshak.

Nobody is quite sure why Congress thought the 180 day rule was important. However, in 1986 keeping your email on your server indefinitely was generally not an option (and even less of an option for, say, voicemail).

ECPA amended the Wiretap Act of 1968.

There are three separate issues regarding email:

US v Councilman

Bradford Councilman ran a website that listed rare books; he also gave email accounts (actually aliases) to booksellers within the domain "" (this might be comparable to giving email aliases to their associated private sellers, or even ebay). However, Councilman examined these dealer emails in order to develop a competitive strategy (these emails would show what rare books were in demand, for example; apparently the real target was

In the case US v Councilman, the government prosecuted Councilman for interception of email in violation of the ECPA/Wiretap Act. Councilman argued that he only examined the email as it was stored on servers temporarily while being routed to its final destination, and that accessing stored documents did not constitute "interception" for the purposes of the Wiretap Act.

The language on interception is quoted above, by the links to the text of the law. It says nothing about the data being "in transit". You might think this would be open-and-shut. But the District Court and a 3-judge panel of the Appellate Court agreed with Councilman's theory. In 2005, however, the First Circuit court ruled en banc that, yes, ECPA in-transit rules did apply to data stored temporarily on disks (filesystems) as well.

Note that the issue here is not government access to electronic communications.

Note also that the status of email as it sits in storage remains contentious.

Email differs technically from voice in that as email is forwarded to its destination the full message sits briefly on various intermediate servers. Phone servers store at most a few bits of a voice stream at a time. The First Circuit ruled very definitively that, despite the appearance that email was being stored, the practical understanding was that it was in transit, and as such was protected. This is a good example of the courts rejecting a "technical" argument for the "big picture"; note, however, that the first two courts to hear the case agreed with the technical argument.

The full First Circuit decision is at

Google is currently being sued over its scanning of gmail to better target advertising. Plaintiffs are claiming that the process violates the ECPA. Google's defense is that users agree to this in its Terms of Service, though Google may also have scanned email from outsiders sent to gmail accounts.

In March 2014, district court judge Lucy Koh ruled that non-Gmail users who sent email to gmail accounts could sue individually, but not as a class action. The legal theory was apparently that there was too much variation between class members as to what Google terms were agreed to. Financially, that makes these cases very difficult. One issue may have been that many non-Gmail users still have Google accounts, or otherwise have agreed to Google's ToS in the past.

However, in a very similar case against Yahoo, Koh ruled on May 27, 2015 that non-Yahoo users who sent email to Yahoo accounts and had it scanned were eligible for class-action status. Part of the issue may have been this sentence from Yahoo's ToS cited by Koh's ruling::

If you consent to this [Additional Terms of Service] and communicate with non-Yahoo users using the Services, you are responsible for notifying those users about this feature.

It would not be unreasonable to imagine that only scanning email sent by ones users would be easy to implement.

United States v Warshak
, civil case, 6th circuit decided June 2007, redecided July 2008, redecided December 2010

This was a case involving government compliance with ECPA. Steven Warshak ran a mail-order operation promoting "Enzyte" for "natural male enhancement." At its peak, annual sales reached $250 million, and the company employed 1500 people.

There were several allegations of fraud: that the physician testimonials for the product were entirely fictitious, that many customers were enrolled a monthly "auto-ship" program without notification, and that the merchant bank accounts were manipulated to make credit-card complaints seem to be a smaller percentage than there actually were. Eventually the government investigated and then prosecuted.

The government got a subpoena order from a US Magistrate asking for his email records. The first request came in October 2004 for Warshak's ISP, NuVox, to retain copies of Warshak's emails. The second request came 180 days later, to turn over any emails that had been archived more than 180 days. In time, all the archived emails were turned over. Warshak and his employees read their email using the POP protocol, which deletes messages from the server as soon as they are read. The feds would have had nothing to subpoena if NuVox hadn't retained copies.

In May 2006 Warshak found out about all this. As the emails were incriminating, Warshak argued that the US needed a warrant, which is much stronger than a subpoena. In 2006 he filed a claim seeking a declaratory judgment that a warrant and not a subpoena was needed. (A declaratory judgment is a court ruling on a procedural matter where there is no actual action ordered.)

Are subpoena rules for email overly broad?
US argument: users of ISPs don't have a reasonable expectation of privacy.

This is clear (post-Smyth) for employer-provided email, though there's no reason to suppose loss of privacy extends to the government.

But what about email from a commercial provider? Are these some kind of business record? Here's an imaginary Yahoo Terms-of-service by Mark Rasch, from :

Because a customer acknowledges that Yahoo! has unlimited access to her e-mail, and because she consents to Yahoo! disclosing her e-mail in response to legal process, compelled disclosure of e-mail from a Yahoo! account does not violate the Fourth Amendment.

The point here is that because Yahoo has access to your email, the government thinks that all your email should be treated just like any other commercial records. You have no "expectation of privacy".

The government argued that this case was like the 1976 US v Miller case, where bank records were found NOT to be protected. However, bank records are pretty clearly different from email. For one thing, under the "transaction" theory of privacy, bank records belong to the bank, as well as to you. Email does not belong, in any sense, to your ISP.

But there's also the issue that ISPs do not just route your email messages, they also store them. Sometimes indefinitely, even after you have read them.

Stored Communications Act, part of ECPA
    electronic communications stored 180 days or less: gov't needs a warrant
    more than 180 days: warrant, subpoena, or court order
Warshak was arguing that the government should need a warrant for ANY of his email.

Back when the ECPA was passed, email existed, but it was not common. The "electronic communications" that Congress had in mind may have been voicemail, for which the 180-day cutoff makes a lot more sense.

At the district court level, Warshak won in his declaratory-judgment quest. (Quote from full 6th circuit decision)

The court reasoned that Warshak likely would succeed on his Fourth Amendment claim because internet users have a reasonable expectation of privacy in e-mails, and because the orders authorized warrantless searches on less than probable cause.

In June 2007 a 3-judge panel of 6th circuit appellate court [Judges Boyce Martin, Martha Daughtry, William Schwarzer (District Court judge sitting in)] again ruled for Warshak. The decision was far-reaching, not specific to the facts at hand; the court issued an injunction forbidding the US government from obtaining emails without a warrant. From the ruling (at

[W]e have little difficulty agreeing with the district court that individuals maintain a reasonable expectation of privacy in e-mails that are stored with, or sent or received through, a commercial ISP. The content of e-mail is something that the user "seeks to preserve as private," and therefore "may be constitutionally protected."

In October 2007 the 6th circuit agreed to an en banc (whole court) review of this ruling.

In January 2008, Warshak's criminal case went to trial. He was convicted the following February.

In July 2008 the full court ruled on the declaratory-judgement request. They decided that the question of whether police needed a warrant to obtain emails was not "ripe": the broad question was not ready to be addressed. The injunction was lifted.

The ripeness doctrine serves to "avoid[] . . . premature adjudication" of legal questions and to prevent courts from "entangling themselves in abstract" debates that may turn out differently in different settings.

That makes sense, but the full court also said some strange things about expectations of privacy:

The answer to that question will turn in part on the expectations of privacy that computer users have in their e-mails -- an inquiry that may well shift over time, that assuredly shifts from internet-service agreement to internet-service agreement and that requires considerable knowledge about ever-evolving technologies.

Some of these service-provider agreements could cast doubt on the validity of § 2703(d) in a given case; others might not. Better, we think, to decide the validity of the statute in the context of a specific internet-service agreement and a specific search and seizure."

In determining the "reasonableness" of searches under the Fourth Amendment and the legitimacy of citizens' expectations of privacy, courts typically look at the "totality of the circumstances,"

Do you think your email privacy from the government should hinge on the agreement you have with your ISP?

(See Eugene Volokh, Traditionally, the courts consider 4th-amendment cases only in concrete contexts and not in the abstract. To be sure, the case as a whole was still at the declaratory-judgment stage; the full court may have felt that the email situation should wait to be decided at the actual trial. But the comments above about the subjective nature of expectations of privacy, and the idea that the terms of service might play a role in this expectation, are unsettling.

Warshak was sentenced in August 2008, to twenty-five years. He appealed to the 6th circuit.

In December 2010, a (different) three-judge panel of the 6th circuit ruled on Warshak's appeal of his conviction ( They ruled that emails were in fact protected under the fourth amendment, and that to the extent that the SCA (above) held otherwise (for emails held longer than 180 days), the SCA was unconstitutional. This time the judges were Damon Keith, Danny Boggs, and David McKeague.

Alas for Warshak, the court also held that Warshak's emails were nonetheless admissible as evidence, because the government had acted in good faith (believing the SCA was good law). The court based its protection of email on the principle that wiretapping of telephones has long been regarded as a fourth-amendment search (that is, requiring a warrant).

As for an ISP's ability to read emails, the court wrote

As an initial matter, it must be observed that the mere ability of a third-party intermediary to access the contents of a communication cannot be sufficient to extinguish a reasonable expectation of privacy.

The case might still be appealed to the full 6th circuit or to the Supreme Court, though it has not happened yet and it is probably too late now.

The 6th Circuit decision appears to make no distinction between emails Warshak sent and those he received

The court compared email to conventional postal mail:

Put another way, trusting a letter to an intermediary does not necessarily defeat a reasonable expectation that the letter will remain private. See Katz , 389 U.S. at 351 ("[W]hat [a person] seeks to preserve as private, even in an area accessible to the public, may be constitutionally protected.")

Given the fundamental similarities between email and traditional forms of communication, it would defy common sense to afford emails lesser Fourth Amendment protection.

The court also considered the Third-Party Doctrine, and rejected it:

But Miller is distinguishable. First, Miller involved simple business records, as opposed to the potentially unlimited variety of "confidential communications" at issue here. See ibid. Second, the bank depositor in Miller conveyed information to the bank so that the bank could put the information to use "in the ordinary course of business."

The second point is arguably the more significant. The bank is the other party to your banking transactions; your email provider is simply an intermediary.

The full 6th circuit did not hear Warshak's appeal en banc, despite the fact that they had earlier ruled in Warshak's favor on the issue of whether a warrant was needed for email. Why? One reason was that the 2010 decision was an actual decision in a criminal case, and not a decision about a declaratory judgement. Another was probably that the 2010 judge panel was apparently held in somewhat greater respect by the bulk of the 6th circuit.

Email has been around for ~20 years. Why has this decision taken so long?

Rep Zoe Lofgren (D-CA) has introduced a bill in Congress (HR 6529 - the ECPA 2.0- Act of 2012) to require warrants for all email searches (including things like Facebook messaging) and all uses of phone GPS location information. It remains to be seen whether this will go anywhere.

One final note: Warshak's company often read a disclaimer to their telephone customers. But:

In the case of Enzyte, sales reps were instructed to lead into the disclosure language by stating that "the product is not a contraceptive nor will it prevent or treat any sexually transmitted disease." According to [Chief Operating Officer James] Teegarden, the thinking was that, "if we started off with a statement about a contraceptive, something other than what it was, that people wouldn't really listen to what we were disclosing to them."

I had no idea those fast-paced disclosures attempted this kind of manipulation.

Another Take

Although the Sixth Circuit did not focus on this, one issue is that the government asked Warshak's ISP NuVox to retain copies of the email. If this had not occurred, there would have been no stored records to subpoena. A more "limited" decision might have focused simply on forbidding this kind of unauthorized archiving, explicitly authorized by §2703(f) of the SCA.

The government's behavior in requesting the archiving does somewhat undercut the idea that they were placing "good-faith reliance" on the SCA.

Are the feds complying?

Not necessarily. But maybe partially.
... since the Sixth Circuit Court of Appeals' 2010 ruling in U.S. v. Warshak, the Justice Department has generally sought court warrants for the content of e-mail messages, but is far less inclined to take that step for non-email files.
New bill proposed that would, in some cases, not require a warrant (though the Sixth Circuit has already found that unconstitutional)
Talks about the "email privacy act".

The Third-Party Doctrine and email

The third-party doctrine (see above) says that you lose any expectation of privacy in anything you disclose to a third party. This is the traditional justification for the government's being able to access any business records about you, including your bank records and cellphone location records.

If you are corresponding via email with a co-conspirator, sometimes the government tries to apply the third-party doctrine to argue that you have no privacy expectation in the emails you sent. This is an extreme position, but one grounded in the idea that the email in your in-box is generally all from other parties.

Judge Tena Campbell, in the Utah case US v Young, decided the case exactly along these lines: defendants lost all privacy expectations in emails sent to other defendants: "[a] sender of an e-mail loses his or her reasonable expectation of privacy in an e-mail that has actually reached the intended recipient." If Alice sends an incriminating email to Bob, and the police seize it from Bob without a warrant, then Bob has no standing to challenge this as it is not his email. Similarly, Alice has no standing as it has been delivered. Judge Campbell explicitly cited US v Warshak:

In Warshak v. United States, the United States Court of Appeals for the Sixth Circuit analogized a search and seizure of e-mails through an ISP to the interception of a letter at a post office [emphasis added - pld]. 631 F.3d 266, 286 (6th Cir. 2010). Here, transmission of all the e-mails and texts was complete before seizure, that is, the e-mails had already arrived at the intended recipient account.

Of course, the Warshak emails were also delivered before seizure; recall that the main thrust of the Warshak case was arguably the 180-day rule that was part of the Stored Communications Act. Judge Campbell was having none of it; a big part of her ruling was to claim that email should work exactly like postal mail in this regard. Once you have received a letter, the sender has lost his or her privacy interest in it. (On the other hand, the feds were interested in Warshak's email file in order to prosecute Warshak, so Judge Campbell's argument would not apply.)

But there is one special circumstance here: the emails in question were sent to defendant Taylor, who originally filed a motion to suppress their use. Taylor then entered into a plea bargain, and withdrew that motion; effectively, he agreed to turn over his received email to the prosecutors. It is Taylor's co-defendants -- the senders of the email -- who were still arguing for suppression; Campbell simply felt it was too late.

Campbell's decision is at,47&case=14463355625153858382&scilh=0

Other arguments about email

The Federal government continues to try to use the argument that your email represents business records of your ISP. Mostly the government has simply ignored the Sixth Circuit ruling, though perhaps most email servers are physically located outside the Sixth Circuit (Kentucky, Michigan, Ohio and Tennessee).

Gag Orders

The Patriot Act (§215) has a gag-order clause:

(d) No person shall disclose to any other person (other than those persons necessary to produce the tangible things under this section) that the Federal Bureau of Investigation has sought or obtained tangible things under this section.

So does §505 of the Patriot Act, covering National Security Letters. So does the ECPA. Are these appropriate?

The short answer is yes, absolutely: in active law-enforcement investigations it is important not to tip off the suspect. The problem is that the gag orders often last indefinitely. The reason nobody knew about all the NSA surveillance before Snowden was simply that none of the gag orders -- some of them many years old -- had expired!

The lawsuit Doe v Ashcroft was filed in 2004 alleging that the gag order involved on the NSL requests for email-account information was unconstitutional. District Court judge Marrero ruled in favor of Doe, but the case was appealed to the Second Circuit. The Second Circuit ruled that gag orders were subject to "strict scrutiny" and returned the case to Judge Marrero as Congress had in the meantime changed the law. In 2010, the FBI agreed to partially lift the gag order, allowing Doe to identify himself as Nicholas Merrill, and to identify the ISP involved as Merrill's Calyx Internet Access. In 2015, Judge Marrero lifted the entire gag order.

California District Court Judge Susan Illston found NSA gag orders unconstitutional in 2013, in a case brought by the EFF on behalf of a telecommunications company ( The Ninth Circuit was going to hear the case, but Congress passed the USA FREEDOM act amending the law. The Ninth Circuit sent the case back to Illston, who withdrew her earlier finding of unconstitutionality. One provision of the new law is that the NSA or FBI needs to defend the gag order in court if the recipient requests it.

Several other US providers (Google and Microsoft among them) have also filed challenges to indefinite gag orders.

Discussion: do you believe people generally should have a right to know when the government accesses their online documents?

Some sites maintain warrant canaries. If they have received no gag orders, they publish that information. If a gag order is received, the information is removed. The principle behind warrant canaries is that "the government cannot order you to lie".
In November 2017 the Justice Department agreed to limit gag orders, most likely in response to a Microsoft lawsuit filed in early 2016. When requesting a gag order as part of a subpoena of customer information, prosecutors must now spell out why they feel this is necessary; an example might be that the government fears the suspect might flee or might destroy other evidence. Gag orders are now to be limited to a maximum of one year, as well. Microsoft continues to lobby to get Congress to limit gag orders to 90 days unless there is a pressing national-security issue.



The saga of Ladar Levison's Lavabit email service deserves a mention here. It was founded in 2004, and abruptly shut down on August 8, 2013. As it later turned out, Levison was served with a warrant (first a subpoena but eventually a warrant) asking him to turn over the TLS keys protecting all communication into and out of Lavabit's servers. The government eventually revealed (perhaps through error) that the target was Edward Snowden; Snowden had released his documents in June 2013. The TLS keys would give the government access to:

Snowden and his correspondents are believed to have deleted emails promptly.

Levison fought with the government for a while, but ultimately decided to shut down the service rather than expose all his users to surveillance.

At the time of the shutdown, Levison was under a gag order forbidding him from discussing the demand for the TLS keys. The gag order was lifted October 13, 2013.

Levison was ordered to appear in court without his attorney; as a third party to the case, Levison did not have an automatic right to an attorney. He was charged with contempt of court at some point, without a hearing. He appealed; the circuit court said they could not hear the case because Levison had not filed an objection with the district court. But he could not do that because there had been no hearing.

Levison recounted his story in May 2014:

Phone Searches

Prior to the Riley decision of the Supreme Court, the situation regarding phone searches was rather unsettled.

In 2010 the US Supreme Court ruled in Ontario [California] v Quon that the City of Ontario had a right to audit pager text messages on pagers issued by the city to police officers. Many messages turned out to be personal, and the senders were disciplined. Officer Jeff Quon sued, arguing that the department had promised that personal messages were ok if officers reimbursed the city for any message-overage charges.

The decision was unanimous that this particular case amounted to a legitimate work-related search, and that Officer Quon should have realized that privacy was not guaranteed here. However, the justices explicitly refused to rule on the broader question of text-message search without a warrant. Justice Kennedy wrote,

The judiciary risks error by elaborating too fully on the Fourth Amendment implications of emerging technology before its role in society has become clear

and alluded to the 1928 Olmstead ruling (reversed in the 1967 Katz decision) that telephone wiretapping did not require a warrant.

In January 2011 the California Supreme Court ruled, in the case of Gregory Diaz, that when someone is arrested at a traffic stop and the police search the vehicle, they may extend the search to at least the text messages on any phones they find. The police rationale here was to tie a suspect to a drug deal an hour before, through SMS messages. Note that the police need only "probable cause" to search a vehicle when they arrest someone (and also in some other situations I am not entirely clear on), though the Supreme Court did rule in Knowles v Iowa that the police may not search a vehicle or its passengers if the stop was for a routine traffic violation.

In March 2012, the Seventh Circuit ruled (in US v Flores-Lopez) that if the police find a cellphone on a person during an arrest, they may look at the phone to find out the phone's number. The police did exactly that when arresting Abel Flores-Lopez, and later subpoenaed the calling records for that number. The decision, written by Judge Richard Posner, likened the phone to a diary found on an arrested person, and established precedent does allow the police to search a diary you have on your person.

In September 2012 a state judge in Rhode Island ruled that the police should have obtained a warrant in the case of the death of six-year-old Marco Nieves. An officer responding to the initial 911 call found a phone on the kitchen counter and proceeded to read the text messages within it. Some of the messages suggested child abuse, and charges were brought against Michael Patino, boyfriend of Marco's mother.

In December 2012, the Massachusetts Supreme Judicial Court ruled that the police may examine the call lists in a phone found on someone they arrested. The case involved the arrest of Demetrius Phifer; after Phifer's arrest the police cross-checked the called-numbers list on his phone and on the phone of a suspected buyer. The police did the phone search after Phifer was brought to the station.

All these cases relate to call lists or to simple SMS messages. What if you also get email on your phone, or if you have a substantial browser or search history, or if your phone has been tracking your location? This is even more unsettled, though recall the Melvin Skinner case.

The Riley Supreme Court case settled most of these phone-search cases.

Eavesdropping on SMS messages ("text messages")

1. They are often transmitted as cleartext.

2. The government is likely to argue that the 4th amendment does not apply to eavesdropping (or to physical examination of the phone, for that matter).

3. They are not 'wire communications', and thus escape the Wiretap Act rule that illegally intercepted messages cannot be used against you.

4. Your local police are not likely to be intercepting SMS messages, but it's always a risk. The ECPA does require a court order.

5. Your wireless provider probably logs and stores all your SMS messages.

More at

And if you are worried about this, get Signal. Or even Whatsapp.