Actionable Intelligence

While listening to the radio this morning, I heard a caller arguing that torture doesn’t work, because by torturing people, you can get people to say anything.

I’ve objected to this argument before, because it implies that torture would be OK if only it were more reliable.

There’s another problem with this line of reasoning, though. Suppose you need someone to say something that isn’t true. Suppose that a single lie, professionally extracted from some anonymous victim, will help you to launch a war, or enact an unconstitutional law, or round up a horde of political enemies, or boost defense spending.

Eventually, you might end up in a situation reminiscent of the subprime mortgage crisis, in which vast fortunes are staked on little more than promises made under duress.

If this ever happened, the problem with the existence of a videotaped torture session might not be its brutality, so much as the insight it’d provide into the process by which “intelligence” is manufactured and passed off as legitimate.

This isn’t an accusation, of course; I’m just thinking out loud. But in purely political terms, this does strike me as the most dangerous possible aspect of allowing torture: it can produce “useful” information when reality can’t.