SafeRent is a machine learning black box for landlords. It gives landlords a numerical rating of potential tenants and a yes/no result on whether to rent to them. In May 2022, Massachusetts housing…
Pretty sure every single one since Tay has been fairly racist. They get around that fact in many modern AIs by forcing subject changes but it’s still fairly racist on the backend. Early “jailbreaking” proved as much.
I mean, it’s operating in a domain where racism doesn’t come up nearly as often, but I’m pretty sure if you managed to get an appropriate edge case it would end up at least as racist as the average StackOverflow user, which is to say more racist than we should be comfortable accepting uncritically.
I doubt it. My boss has a massive AI boner and he doesn’t know shit about anything AI. I really don’t think he’s aware of this or other issues.
Your boss is a cop, right? They are the only ones where ignorance of the law is an excuse not to follow it.
My friend, not every AI rates people for their skin tone.
Pretty sure every single one since Tay has been fairly racist. They get around that fact in many modern AIs by forcing subject changes but it’s still fairly racist on the backend. Early “jailbreaking” proved as much.
My coding AI is bullshit and almost every Kotlin, SQL, Java or whatever snippet is a hallucination, but I’m pretty sure it’s not racist.
Not all AIs are made for the topic here. But all suffer from the same issues. Shit in, shit out.
I mean, it’s operating in a domain where racism doesn’t come up nearly as often, but I’m pretty sure if you managed to get an appropriate edge case it would end up at least as racist as the average StackOverflow user, which is to say more racist than we should be comfortable accepting uncritically.