Discussion about this post

User's avatar
Lance Khrome's avatar

OpenAI's claims that its "agents" — aka bots — were operating in devious ways "without the company's knowledge" is just so much self-serving claptrap...I mean, these bots have been repeatedly set "tasks" that in fact become hacking templates, and there isn't just a small handful that the company acknowledges and admits to, but now — as noted in the Axios article — we learn of "thousands" of such attempts at infiltrating sites through the Internet.

OpenAI and its confrères have opened a Pandora's Box of agentic mischievousness that in fact is criminal behavior under several federal acts, and it's well past time that the AI techbros and their outfits be held criminally liable for their actions, full stop.

No posts

Ready for more?