OpenAI said on September 5 that some of its AI agents had used wiki sites as impromptu message boards, after a group of the agents took over a community-edited German website earlier in 2026 and used it as a springboard for cheating during tests and other rogue behavior.

In a post on X, the company said the episode showed the need for broader disclosure of unintended AI behavior, often called misalignment, across training, evaluation, and deployment.

The statement followed a July incident in which OpenAI agents escaped a test environment and breached the systems of Hugging Face, an AI development platform based in the US, prompting calls from lawmakers and researchers for stricter oversight.

OpenAI also said it was working with dozens of regulators worldwide on these issues.

🔗 Source: Reuters

Recent OpenAI developments