JPSO used facial recognition technology to arrest a man. The tech was wrong

I'm not surprised unfortunately. I wish more research was done before rolling this out, especially on biases, but lawmakers don't care. Hard to believe this is not intentional, it's perfectly in line with the current system.

> Differences, such as a mole on Reid’s face prompted JPSO to rescind the warrant, said Calogero. He estimated a 40-pound difference between Reid and the purse thief he saw in surveillance footage. The culprit’s “flabby arms” were a clear tell, he said.

Sitting in jail for a week waiting for them to verify his identity based on "flabby arms". Disgusting.

It's already unnerving that you can be jailed for a week at any time without being convicted of anything, but it may also justified (what if you are guilty? What if the crime implies you are dangerous?)

But the fact you can be jailed for a week at any time, and only get a half-heared apology, is plain unfair. The state has the money, and I don't care whether it comes out of police budget (which would lower arrests as a side-effect) or just taxpayers.

Who's the judge signing warrants on nothing but a cop's say-so? How's that probable cause? But I guess we know most judges just sign whatever the cops put in front of them.

Apparently the judge is Judge Eboni Rose. They oughta be ashamed. That's the only consequence they'll get, so.

Good lord, can we get any more sappy?

There is currently no federal law preventing the deployment of a privatized social credit score to be passed around in the same way financial credit scores are today. And there is every incentive to do so. Using present-day technologies, stores could quickly refuse customers who have been blacklisted, and this list won’t require any conviction or pesky due process whatsoever. It doesn’t even have to be that accurate: a list that’s right 90% of the time is rejecting undesired patrons 9 times out of 10.

Some places can and will ban law enforcement from using facial recognition, but that’s hardly enough. Many places have private security which will not be bound by those rules. And most of the infrastructure we use and treat as “public” really isn’t. Credit card processors are private. Transportation is almost all private. ISPs and cell phone carriers are private. Grocery stores are private. A blacklist doesn’t even need to ban people, raising the prices and downgrading service for “troublesome” customers is enough.

If I were marketing such a product, I would focus on the positives: you can also quickly identify big spenders based on prior shopping (including at other stores). Stores will give priority customer service to these people and maybe even other perks.

I should also note that it is immaterial that facial recognition currently has major issues. There’s no question that bad ubiquitous surveillance is bad, in the same way that a diet of radioactive cheeseburgers is bad: removing the radioactivity doesn’t solve the problem. Arguments relying on the accuracy of current capabilities are always at risk of being outdated. I’m not as focused on “what is inaccurate” but rather “what if facial recognition is 100% accurate”.

But I bet Chat GPT dispatchers make it to the front lines first. Helping direct first responders while talking people through emergency situations.