Tuesday, October 25, 2016

Systemic risk in leak-based electioneering

I recently interjected into the following twitter conversation, wherein @thegrugq asked if media self-censorship in the face of state-sponsored election meddling was warranted. Mr. Greenwald's answer, was a resounding, "No."
Mr. Greenwald went on to publish a related, thoughtful, long-form article in The Intercept. If you've not read that, start there and come back.

Let's presume that Mr. Greenwald's lemma #5 is satisfied, and that a particular leak of dubious legality has much material that is expressly in the public interest. Let's also assume that in a given election cycle, the various candidates have dirt to be uncovered that would also satisfy the bar of public interest. Then, given
  1. Election cycles have a limited duration.
  2. In any specific time period, the media has a whole has limited resources to investigate, vet, and publish.
  3. It costs less to validate leaked documents' authenticity than to uncover the same problem through direct investigation.
it follows that the most efficient media strategy, both in terms of cost and journalist-time, is to mine leaks and report on them before falling back on investigation.

Whereas Mr. Greenwald's commentary is a great decision criteria for publishing any one piece of revealed information, it does not address the concern of what to choose to follow-up on among multiple leaks, or what to do when you have leaks about one candidate, and investigative journalism to do on the other. Obviously reporting on the things you know in the public interest is important, but if the media on the whole satisfy themselves at the feeding trough of the leak, then the media will be complicit in differential coverage of the candidates. This model relies on partisan or contrarian journalists leaving said feed trough because their interests aren't simply costs or efficiency.

I think in the long run, information can find its way out: leaks run out of new reportable data and people forced to look elsewhere. But in the scope of a campaign, that long run may be well after election day, and leaks can serve as the smoke and mirrors distracting the electorate.

Furthermore, it follows that anyone wishing to influence the election is given incentive to, by any means possible, dump as much as they can get about their opponent and hope that as much dirt as possible passes the long-tail of each journalist exercising their own filter on what counts as public interest. I'm generally pro-transparency, but setting a high bounty on exfiltrating private data to help nudge elections toward having friendlier candidates in office seems like an inherent risk in this model of journalism.