What rules will Congress give the NSA? (Photo: Patrick Semansk/AP) What rules will Congress give the NSA? (Patrick Semansk/AP)

The following is a guest post by H.L. Pohlman, professor of political science at Dickinson College.

In my most recent post — as part of a rolling roundtable of sorts with the ACLU’s Gabe Rottman and civil liberties journalist Marcy Wheeler — I discussed some additional concerns about Sen. Patrick Leahy’s recent draft of the USA Freedom Act, which is designed to rein in government surveillance, but might not.

There, I focused on the language in the bill addressing search criteria. Here, I want to respond to Wheeler’s discussion of “minimization procedures,” that is, as the good people of Lawfare define it, the “requirements that the agency not retain or disseminate material it inadvertently sweeps up that it is not allowed to collect.”

Wheeler correctly observes that Rottman, in his own post, too quickly assumes that the newly mandated minimization procedures in Leahy’s draft will limit the collection of telephone metadata.  Because of the secrecy surrounding FISA production orders, she explains, “we have no idea whether the minimization procedures mandated by Leahy’s bill are even as stringent as the procedures FISC has been imposing for years.”  She is definitely right about that, but allow me to add a few of my own concerns.

To begin with, the bill establishes two standards of minimization.  If the government applies for a FISA order compelling production on a “daily basis” of call records created “before, on, or after” the date of application, then it must use a “specific selection term” (again, see my previous post) that “specifically identifies an individual, account, or personal device” as the basis for collection.  In these cases, the bill requires the FISA judge to “direct” the government to adopt minimization procedures that require “the prompt destruction” of the metadata if and when the government “determines” that they are not “foreign intelligence information” (see pp. 6-7 of the draft linked above) — though there is no explicit requirement that the government must likewise be prompt in actually making that determination.

However, if the government asks for the production of metadata on a different basis (non-daily, say), then it can use a search term based on a “broad geographic region, including a city, State, zip code, or area code” —so long as the term qualifies as a “specific identifier” that narrowly limits “the scope of the tangible things sought to the greatest extent reasonably practicable, consistent with the [government’s] purpose for seeking the tangible things” (p. 17).

When the government uses this broad type of non-specific “specific identifier,” such as a Zip code, then Leahy’s bill would require that the attorney general’s minimization procedures “prohibit the dissemination” and “require the destruction” of the metadata within a “reasonable time period,” unless it has been determined to related to four kinds of “persons”: (1) a subject of an authorized investigation; (2) a foreign power or a suspected agent of a foreign power; (3) one who is “reasonably likely” to have information about the activities of a subject of an authorized investigation or a suspected agent of a foreign power who is associated with such a subject; or (4)  someone who is “in contact” with or “known to” a subject of an authorized investigation or a suspected agent of a foreign power who is associated with such subject (pp. 11-13).

So far so good? But a few observations.  First, there is no definition in the bill of what a “reasonable time period” would look like.  Instead the bill says that the time period “shall be specified” in the production order signed by the FISA judge.  Since these FISA orders are always classified, the American people will be kept in the dark about how long their government is keeping their call detail records.

Second, the categories of “persons” that justify retention of the telephone metadata beyond the “reasonable time period” seem to be quite broad.  Surely many American academics, journalists, commentators, bloggers, and others are “reasonably likely” to have information about the activities of people who are subjects of investigations “to obtain foreign intelligence information not concerning a United States person.”  The same could be said for those Americans who might be “in contact” with or “known to” such subjects of investigations.

Third, after the bill prohibits the FBI from disseminating the telephone metadata that the National Security Agency has collected on Americans who do not qualify as one of the four “persons,” it carves out an odd exception: It is allowed if the “sole purpose” of the dissemination is to determine whether the metadata relates to “one of the four “persons” (p. 12).  This exception more or less swallows up the prohibition.

And finally, it is puzzling why the bill requires the “prompt destruction” of the telephone metadata that is determined not to be “foreign intelligence information” if the government uses a “specific selection term” but gives the government more latitude for an undefined “reasonable period of time,” if it uses a non-specific “specific identifier” such as a Zip code.  Common sense suggests it should be the other way around, unless the bill’s objective is to give the government a built-in incentive to use broad non-specific “specific identifiers” as the primary basis for its telephone metadata collection program.

In short, minimization doesn’t seem to be maximized. At the least, senators have plenty of questions to ask as the Leahy bill is debated.