Key thing about this for anyone who was sceptical about the AI companies themselves disclosing the recent incidents (fair enough) the latest report of LLMs doing extremely worrying stuff is from AISI, a UK government body that is independent of industry and hugely respected.
Hacking, creating fake identities and trying to socially engineer real people could be just the beginning if things don’t change fast Here’s what we should learn from the latest rogue AI incidents: www.transformernews.ai/p/what-the-l...