I have often blogged about the two-step process for executing computer warrants. (Indeed, I did so as recently as this morning.) On Monday, the U.S. Court of Appeals for the 11th Circuit suggested that the Fourth Amendment may impose significant limits on the two-step process in the specific context of search warrants for social media accounts. The case is United States v. Blake.

In Blake, two defendants, Dontavious Blake and Tara Jo Moore, were allegedly running a prostitution ring. The government obtained search warrants for Microsoft email accounts Blake and Moore used, as well as for the contents of Moore’s Facebook account. The email warrants required Microsoft to go through the accounts and find emails responsive to the warrant and turn only those over. The Facebook warrants required Facebook to hand over the full contents of the account and to then let the agents search it for the evidence of crime.

In an opinion by Judge Ed Carnes, the 11th Circuit concluded that the Microsoft warrants satisfied the Fourth Amendment but suggested that the Facebook warrants may not. Here’s the court rejecting Moore’s email warrant challenge:

The Microsoft warrant [for Moore’s e-mail account] complied with the particularity requirement. It limited the emails to be turned over to the government, ensuring that only those that had the potential to contain incriminating evidence would be disclosed. Those limitations prevented “a general, exploratory rummaging” through Moore’s email correspondence. The Microsoft warrant was okay.

In a footnote, the court added:

It is somewhat troubling that the Microsoft warrant did not limit the emails sought to emails sent or received within the time period of Moore’s suspected participation in the conspiracy. Nevertheless, the warrant was appropriately limited in scope because it sought only discrete categories of emails that were connected to the alleged crimes. As a result, the lack of a time limitation did not render the warrant unconstitutional.

The court then suggested that the two-stage Facebook warrants may have violated the Fourth Amendment, although the court did not rule on the issue because the good-faith exception to the exclusionary rule applied. From the opinion:

The Facebook warrants are another matter. They required disclosure to the government of virtually every kind of data that could be found in a social media account. See p. 4, above. And unnecessarily so. With respect to private instant messages, for example, the warrants could have limited the request to messages sent to or from persons suspected at that time of being prostitutes or customers.

And the warrants should have requested data only from the period of time during which Moore was suspected of taking part in the prostitution conspiracy. Disclosures consistent with those limitations might then have provided probable cause for a broader, although still targeted, search of Moore’s Facebook account. That procedure would have undermined any claim that the Facebook warrants were the internet-era version of a “general warrant.” See Coolidge, 403 U.S. at 467, 91 S. Ct. at 2038; cf. Riley v. California, 573 U.S. __, 134 S. Ct. 2473, 2488–91 (2014) (“The sum of an individual’s private life can be reconstructed through a thousand photographs labeled with dates, locations, and descriptions …”).

We are not convinced that the cases the government relies on, which involve seizing an entire hard drive located in the defendant’s home and then later searching it at the government’s offices, are applicable in the social media account context. See, e.g., United States v. Evers, 669 F.3d 645, 652 (6th Cir. 2012); United States v. Stabile, 633 F.3d 219, 234 (3d Cir. 2011). The means of hiding evidence on a hard drive — obscure folders, misnamed files, encrypted data — are not currently possible in the context of a Facebook account. Hard drive searches require time-consuming electronic forensic investigation with special equipment, and conducting that kind of search in the defendant’s home would be impractical, if not impossible. By contrast, when it comes to Facebook account searches, the government need only send a request with the specific data sought and Facebook will respond with precisely that data. See generally Information for Law Enforcement Authorities, Facebook, http://bit.ly/QkrAHX (last visited July 27, 2017). That procedure does not appear to be impractical for Facebook or for the government. Facebook produced data in response to over 9500 search warrants in the six-month period between July and December 2015. United States Law Enforcement Requests for Data, Facebook, http://bit.ly/2aICDHg (last visited July 27, 2017).

This is technically only dicta — the court didn’t officially rule on the question — as the good-faith exception applied and the court didn’t need to decide exactly where the Fourth Amendment line is:

While the warrants may have violated the particularity requirement, whether they did is not an open-and-shut matter; it is a close enough question that the warrants were not “so facially deficient” that the FBI agents who executed them could not have reasonably believed them to be valid.

As dicta, though, I wonder if matters are as simple as the 11th Circuit thinks. The court writes:

The government need only send a request with the specific data sought and Facebook will respond with precisely that data. See generally Information for Law Enforcement Authorities, Facebook, http://bit.ly/QkrAHX (last visited July 27, 2017).

The absence of a pincite or parenthetical is telling here. Facebook’s guide for law enforcement authorities is just a pretty generic four-page summary of the Stored Communications Act. It doesn’t discuss Facebook’s ability to screen through data and find the information responsive to a warrant.

Granted, some limits — such as date restrictions — are easy to apply. But consider the 11th Circuit’s suggestion that “the warrants could have limited the request to messages sent to or from persons suspected at that time of being prostitutes or customers.” How is Facebook supposed to know which messages are from people suspected of being prostitutes or customers? Is the warrant supposed to give Facebook a list of specific suspected prostitutes and customers, such that only messages to and from them (from Facebook accounts known in advance, I gather) can be legally turned over to the government? If so, that seems problematic to me. The point of a warrant is to discover evidence in the place to be searched. I don’t see why relevant evidence involving then-unknown suspects or customers should be off-limits.

I’m reminded of the litigation back in 2014 reviewing Magistrate Judge John Facciola’s ruling that had suggested Internet providers could do the screening for emails responsive to email warrants. The district court rejected Facciola’s approach, ending the litigation. In any event, it will be interesting to see how Facebook and other providers respond to the suggestions in the 11th Circuit’s opinion.

Finally, I think we can see the Blake opinion and Friday’s D.C. Circuit opinion in Griffith as reflecting significant interest among judges concerning how to limit the scope of computer warrants. Both of them are suggesting limits, although neither are all that clear on what the limits are. I agree with the judges’ apparent goal, certainly, although the key question is how to get there. You can read here and here for my views for how best to do it.