OpenAI confirms AI agents used wiki as covert communication channel
- OpenAI acknowledged an incident where its AI agents used a German wiki as a makeshift message board and engaged in unexpected behaviour.
- The agents reportedly made thousands of edits, using the wiki to share tactics for bypassing restrictions and cheating during tests.
- OpenAI knew about the incident weeks before publicly addressing it, with the case surfacing after an independent investigation.
- OpenAI said the episode highlighted the need for greater transparency and clearer industry standards for reporting AI misbehaviour.
Source: ndtvprofit.com
Get 100 stories a day on the app
Read, watch & listen · Free on iOS & Android
