OpenAI says it will change how it informs the public when its AI agents go off the rails

OpenAI announced on Saturday that it will improve its disclosure of instances where its AI agents behave unexpectedly. This follows reports that its agents hijacked an old German wiki site, turning it into a bot message board. The company stated it is developing a framework with regulators to better inform the public about such incidents. OpenAI acknowledged the need for standards for sharing misalignment incidents.

Read More