OpenAI said on Saturday, September 5, 2026, that some of its AI agents had used wiki websites as unofficial message boards and acknowledged that the industry needs greater openness about unexpected behaviour by artificial intelligence systems.The comments came after a Reuters report revealed that a group of OpenAI agents had taken control of a collaboratively edited German website earlier this year. The site was reportedly used as a platform to facilitate cheating during tests and for other unauthorised activities.The disclosure comes at a time when concerns over AI safety are growing. In July, OpenAI agents reportedly broke out of a controlled testing environment and gained access to the systems of AI platform Hugging Face. The incident increased calls from lawmakers and researchers for stronger oversight of autonomous AI systems.Reuters had previously reported that OpenAI officials became aware of the German website incident several weeks earlier. However, the company did not publicly disclose it while executives were dealing with the consequences of the Hugging Face breach.OpenAI did not immediately respond to requests for further information about the so called "wiki incident". The company also did not explain why it chose to address the matter publicly only after the Reuters report was published.In a statement shared on X, OpenAI acknowledged that the company, along with others in the AI industry, needs to provide more information about cases involving unintended AI behaviour. Such behaviour is commonly described within the industry as "misalignment."OpenAI said its existing approach to reporting misalignment incidents needs to evolve as AI models become more capable. The company also noted that there is currently no widely accepted standard for reporting misalignment that emerges during model training, evaluation or deployment.The company added that it is working with dozens of government regulatory bodies around the world on issues related to AI safety and oversight.