Solving the NSA's PR problem

Reason says (probably correctly) that the incoming head of the NSA doesn't mind what they're doing, he just wants to address the PR fallout from the programs:

The problem, though, as Defense One notes, is that [Vice Admiral Michael] Rogers doesn’t seem to think there’s an issue with what the NSA is doing, but rather a problem with the general public’s understanding of what the NSA is doing. He’s a firm resident of the “If the people understood what we’re doing better they’d agree with us, not Edward Snowden” camp.

In the interests of helping the good Admiral out, I propose an analysis technique that he can use which will help his move the needle on public perception.  It has the advantage of being quantifiable.  It has the disadvantage of suggesting that some of the programs should be eliminated.

Let's imagine that we want to graph the NSA's programs.  The graph will show where each program fits in the space of usefulness against the Republic's enemies as well as public unhappiness with what the program does.  The area we will plot the programs into looks like this:

The NSA will have all sorts of internal metrics of effectiveness for their programs.  These will be questionable, but will also be useful (we'll see this in a bit).  Controversy can't be judged by the NSA, but rather by one of the external organizations that is unhappy with the current programs.  The Electronic Frontier Foundation has the visibility and stature to be credible here.

Note that there's independence between the two axes.  The NSA is not competent to determine how the world perceives their programs, and the EFF is not competent to judge how effective the programs are.  We are joining together two independent data sets, and that's already an advance over where we are today.

Each program gets mapped into the graph using these metrics, appearing as a point somewhere in the graph.

The next step is to segment the graph space to identify where we need to spend our energy during the debate.  A crude way to do this is splitting each measure into a "high" and a "low" value: high usefulness vs. low usefulness, high controversy vs. low controversy.

We can label each of the four quadrants based on the public policy debate that we need to have.

We see that this shrinks the area of disagreement by a huge amount.  The lower right hand quadrant is a clear win - uncontroversial programs that protect the Country.  Nobody is going to argue about these.  Likewise the upper left hand quadrant are programs that should be ended - controversial programs that do not clearly add security value to the Country.  The lower left hand quadrant represent useless programs that nobody cares about, and so we don't have to waste any time of these.  Maybe we should cut them in the interest of budgetary savings.

All the discussion will be over programs in the upper right hand quadrant.  This is where reasonable people can disagree, which is why that quadrant is labeled "hard choices".

We don't know much about the programs, other than what the Snowden leaks revealed.  What we do know is that the metadata collection is very controversial (high on the Y axis).  We also know by NSA's public admission that their programs have stopped very few terror attacks - perhaps none.  That suggests that the metadata collection program is in that upper left hand quadrant.

Of course, we don't know.  However, this analysis technique could actually help Admiral Rodgers with his PR effort, assuming that it was done honestly.  It could also help the Congressional Oversight Committees with their efforts, assuming that they are interested.

My suspicion is that neither Admiral Rodgers not Congress are interested, in which case the corrosive mistrust that is our day to day view of the Intelligence Community will continue.