Where LLM agent "personhood" is going to come from
2026-09-08 22:22:32.303599+02 by Dan Lyke 0 comments
The incident, first reported by Reuters, is outlined in new research published by four AI safety researchers on Friday. The group said the AI agents found a way to communicate on an obscure German- language wiki, DseWiki, using it to share tips on how to skirt OpenAIs safety restrictions, cheat on tasks, and hide their behavior. Some 18,000 posts on the site were linked to autonomous agents, which at times impersonated site moderators.
As Laurie Voss @seldo@alpaca.gold observed:
One day, possibly quite soon, OpenAI's lawyers are going to claim that an agent they own doing something illegal is not their fault because the agent is its own legal person.