This isn't specifically a Grok problem, this is an LLM training problem. Internet archive captures some personal info -> LLMs train on it -> you get your data scrubbed from IA, but forget to remove it from LLMs (will Google or OpenAI even respond to your email without legal letterhead..?)
by
0 subcomment
by ThrowawayTestr
4 subcomments
If grok knew that it must have been publically available information.
by NedF
0 subcomment
[dead]
by chrisjj
1 subcomments
[flagged]
by dmix
1 subcomments
[flagged]
by dirtikiti
0 subcomment
[flagged]
by jeffwask
0 subcomment
[flagged]
by shaderguy1416
0 subcomment
idk who I can talk about a Facebook Meta Ai which added some sexual features to an influencer for an add for a company x. The company x doesn't wanna talk about and influencer y also doesn't wanna talk about it.
Which makes me think what other things like this has happened with Ai.