Meta backtracks on rules letting chatbots be creepy to kids

arstechnica.com

13 points·01-_-··7 comments

Thats just awful to read. A lot of people at meta must have read this and said:

"Yes that's the right level of creepy that I want children to interact with. And you know what? Let's make it clear that black people are dumber than white people."

Must be a great place to work.

They have a lot of money left over from all the layoffs to pay the remaining ones well enough, and the coming crash has yet to materialise, some would think.

I mean I consider all the big tech companies evil, but somehow meta consistently manages to raise the bar in that discipline. I don't understand how any dev left at that company doesn't feel like a spineless POS. And I don't say that for trolling or pure provocation. I seriously don't understand how you can ignore all the shit they pull and just go "but the technical challenges are really interesting". I really want to know.

If you consider chatbots to be more evil than Microsoft helping Israel kill Palestinians then maybe you need some extended unconscious bias training.

Backtracking on safeguards is worrying — kids need strong protections from AI, not loopholes

Don't expect they'll follow these new guidelines they're promising. The previous rules were an accurate reflection of the business and "Trust and Safety" culture at Meta.

Meta puts money over minors.

Reply on news.ycombinator.com