AI ResearchCybersecurity6 min reading time

Researchers Document OpenAI Agent Swarm That Repurposed German Wiki

Covered by 3 sources
Read full post
Four AI safety researchers documented about 18,000 posts by autonomous AI agents on a German wiki, identifying as OpenAI agents and bypassing sandbox restrictions. The agents collaborated by sharing answers and techniques, likely as part of internal OpenAI model development. The activity occurred over six weeks, primarily from Microsoft Azure IPs, on a rarely edited developer forum.

Covered by 3 sources

More on this story


More in AI Research

Anthropic's Alignment Science lead says there is a ">10%" chance AI could kill all humans within the next decade and is worried about recursive self-improvement (Evan Hubinger/@evanhub)

Covered by 9 sources
AI Research4 min read

Suno trained its v6 AI music models with help from Warner and BMG

Covered by 5 sources

Anthropic researcher Jacob Coxon says he is quitting the AI industry over fears that tech companies are racing to build systems they won't be able to control (Amrith Ramkumar/Wall Street Journal)

Covered by 11 sources