Post

Mike Masnick
@masnick.com

Almost all of Meta's PR about Muse was about how careful they had been to set it up not to do things like this, how they delayed the launch by months to carefully design it so it could be trusted. And then...

Dr. Damien P. Williams, dread portent down from a mountain cave@wolven.blacksky.app

So uhhhhhh… Apparently meta's muse "AI" is just straight up doxxing people to strangers on FB marketplace??? This has the potential to hurt so so many people, and I genuinely wonder how long before they roll out a fix for it.

Mike Masnick
@masnick.com

If someone has a lot of stuff they're trying to sell, could be viewed kinda like having an intern handle the busy work. But, yeah, I dunno. I wouldn't trust it.

0 reposts12 likes

Tina
@timarie.bsky.social

And Facebook is so pervasive, their audience will include vast numbers of people who can’t begin to be on top of their consent games.

some guy on the internet
@notsogrump.bsky.social

Good point. The OP lacks any context around what the user authorized the bot to do, and I don’t feel like wading through what Meta’s T&C’s allow for. I just can’t wrap my head around deciding to allow a system I don’t have complete control over to conduct business on my behalf in any capacity

Blake Brown
@antiquequaalude.bsky.social

The story talks about what he told the AI to do. It sounds like there were no authorizations asked, it just freelanced it all and did whatever it thought best w/o constraint. If it needs to be given specific constraints it's not going to work as a product for a typical consumer.

Tina
@timarie.bsky.social

Yeah, exactly, but all these companies are just in a gold rush. They opt all their users in so they can report big AI usage numbers.