Opposite Day: Cops Should Be Robots

[Opposite Day caveat: This post started out as a genuine attempt to play with an idea. I was not at all sure I wasn’t willing to give some fairly serious consideration to actually adopting it as a position to advocate. As I wrote it, I realized that Opposite Day was coming up, and I thought it might fit into that scene pretty well, since I also had some strong doubts. Now I think I made the right decision about that. I.e., now I don’t think I think this is such great idea anymore. But I’m interested in what others think about it.]

Perhaps we should make it a priority to develop robots that could perform many of the roles of beat-walking and other street-level police officers, and seek to largely replace human street police with them as soon as we can.

The two outstanding problems relating to excessive police violence, it seems to me¹, are, first, that police use as much force as they feel is necessary to maintain peace and order and enforce the laws while exposing themselves to no more than some certain amount of physical risk to themselves. Meaning, they use more force than is necessary to maintain peace and order and enforce the laws out of fear for their own well-being. We, the public, though, would prefer that they use only strictly as much force as is really necessary to maintain peace and order and enforce the laws, and no more.

Second, police, like nearly all humans, carry around with them unconscious and reflexive biases in how they think of people they encounter with various visible characteristics. The evidence is fairly clear that this results in people of color suffering much more violence at the hands of, well, everyone, but certainly police than white people do. The better we come to understand natural human implicit racial (and other) bias, to me the more unlikely it seems we’ll ever be able to really make significant progress in combatting it. That may be excessively pessimistic, but that’s how it looks to me.

SO, here’s the idea. If the bulk of street policing could ever be accomplished by robots, it should be possible to program the robots to have essentially zero regard for self-safety, allowing them to use only as much force as necessary to maintain peace and order and enforce the laws. This seems like it should allow for less force to be used in the course of making arrests and generally interacting with the populace.

Also, the better we come to understand humans’ biases relating to superficial human traits in others, and the better we come to understand computer learning, perhaps the better we’d be able to exclude it from thinking and reflexes in (eventual) non-human brains. Robot cops should (Obviously this entire item is self-consciously, or perhaps just hypothetically, wildly optimistic about prospects for technological progress in this are) ideally be able to operate with essentially no racial or other common human bias regarding external human traits. Hopefully in the future we will be able to program robots to treat human subjects with the kind of equality that we can’t seem to muster ourselves.

Clearly there would be drawbacks. I expect robot capabilities in community policing would stay considerably behind those of human police for years and years after robot police became an initially vibe technology. But what robot police could do would be to allow humans to radically redefine the role of the human police officer, such that she might be redefined as not a police officer, but as an (unarmed) community-aid agent, who by definition of the job does not enter situations of significant danger, and is not authorized to use any more force in any circumstance than any other civilian human.

More ominously, it’s hard not to fear that accountability for inevitable accidents and incidents of unjustified violence would be even more fractured than what we have today. To a degree, however, this makes me reflect on what it is we’re asking from beat cops while also reflecting on the inadequacy of our present system of accountability for them.

Police unions would obviously oppose this because of the resulting loss of jobs, authority, and outright physical power in society. They would probably successfully defeat it outright, and if not that then forestall it for decades, if it ever became a realistic possibility. But notwithstanding technical and social-political obstacles, and notwithstanding many extremely reasonable concerns about effects and accountability, it’s not clear to me that this isn’t a blue-sky possibility worth evaluating, whose technical plausibility seems something less than completely remote.

What is your reaction to this idea?

Notes:

1. Asking you to work with me here. Presumably the dynamics contributing to excessive police violence are numerous and complex.

(Featured image associated with this post on OT main page from http://www.davidicke.com/forum/archive/index.php/t-204690-p-12.html.)