Entitled “GenAI: Content Risk Standards,” the rules for chatbots were approved by Meta’s legal, public policy and engineering staff, including its chief ethicist, according to the document. Running to more than 200 pages....
They probably just asked AI to summarize it for them and never read it.
Neil Young is very vocal about leaving platforms he somehow finds his way back to when people aren’t paying attention. I don’t disagree in this case but it’s not exactly a barometer of artists in general.
Pornography and violence or whatever are being age gated in the corporate web but the biggest players get to sell sex chats to kids. Pedophilia is legal as long as there is ad revenue to be made.
When I read the headlines about this initially yesterday I was like ‘oh, is this a conservative thing where they’re mad about kids asking the chat bot questions about sex?’
So I read the examples, and HOLY SHIT it is so not that! There is no way to describe it except erotic roleplay that is explicitly about children.
Given the prompt examples I am skeptical that any child would write them, they sound more like shit they grabbed off some pedo ERP site.
It sounds like some idiot working at Meta used an LLM to produce that internal document. Those "acceptable things to say" quotes are just stupid. They have to be quotes an LLM added to the document, unnoticed. There's no way a human wrote them, or even copied and pasted into the doc, and said, "These are acceptable." I refuse to believe that.
They probably just asked AI to summarize it for them and never read it.