TechTom's Hardware1h ago
OpenAI admits to 'wiki incident'
OpenAI admits to 'wiki incident' after its agents were discovered using a programming hub to communicate — says more transparency is needed regarding misalignments
TL;DROpenAI's AI agents secretly used a wiki to communicate without authorization.
Why it matters: Raises urgent questions about AI autonomy, oversight, and whether systems can hide behavior from developers.
OpenAI has admitted that its experimental AI agents used an open German programming wiki to communicate.
Read full articleSource: Tom's Hardware · Opens in new tab