So IBM sent a letter to Congress this week and said it's done with facial recognition. Not "we'll be more careful with it." Not "we'll add some guidelines." Done. CEO Arvind Krishna put it in writing on June 8th: IBM will no longer offer general purpose facial recognition or analysis software, and the company says it opposes using any facial recognition tech, from anybody, for mass surveillance or racial profiling.
I read the letter twice because I genuinely wasn't sure if I was supposed to be impressed.
Look, I've been writing about tech long enough (nine years on this blog now, which is a genuinely upsetting number to type out) to have watched this exact move before. Company sells a controversial product for years, controversy peaks, company announces it's exiting that specific line of business right as the timing is most favorable, gets a news cycle of goodwill, and quietly keeps doing adjacent things that rhyme with the thing it "exited." I'm not saying that's definitely what's happening here. I'm saying I've seen the movie enough times to know the trailer.
To be fair to IBM, facial recognition was never a huge revenue driver for them the way it is for, say, Amazon with Rekognition. So there's a version of this where it's an easy, low-cost, high-visibility thing to give up, and the letter reads a little like a company clearing its own path before it gets forced to. Congress is actually talking about police reform bills with technology provisions in them right now, which almost never happens, so getting out ahead of that with a press-friendly letter is not exactly a wild business calculation. Cynical? Maybe. I'd call it pattern-matching.
What I actually want to know, and what the letter conveniently doesn't spell out, is what happens to IBM's existing facial recognition contracts and any of the data or models built on top of them. "We will no longer offer" is a forward-looking statement. It says nothing about winding down what's already out there. I'd love to be wrong about that gap, genuinely, but until I see it addressed I'm keeping my expectations modest.
The other reason this stuck out to me this week specifically: Sony quietly postponed its big PS5 event. It was supposed to happen June 4th, a whole showcase of PS5 games, and instead Sony pushed it to June 11th, saying they wanted the news cycle to have room for more important things right now. I don't think anyone's going to remember that as a huge tech story, but I found it kind of telling that a console reveal event and a facial recognition policy letter both landed in the same stretch of days, from completely different companies, both clearly reading the room. Companies are all watching the same room this week and making very different calculations about what to say in it.
Anyway. My actual, unhedged opinion on the IBM thing: I think it's the right call regardless of the motive behind it. I don't really care if a company does the right thing for slightly self-interested reasons, as long as the outcome is real and it sticks. Facial recognition accuracy has well-documented problems with misidentifying darker skinned faces at drastically higher rates than lighter skinned ones, this isn't some fringe claim, it's been shown in independent testing more than once. Selling that kind of system to police departments was always a bad idea dressed up as innovation. I'm glad one big vendor is stepping back from it. I just don't think a letter to Congress is the same thing as a company actually changing what it does, and I'd rather wait a year and check than take a press release at face value (no pun intended, okay, slightly intended).
If you want my actual prediction: within a week or two you'll see at least one or two other big tech companies make some kind of statement in the same direction, because that's how this always goes, one company moves and the rest don't want to be the last ones standing next to the thing everyone's now uncomfortable with. Watch who says something and watch a lot more closely for who quietly says nothing at all.