OpenAI, Google, Anthropic Plot Joint AI Oversight Body

OpenAI, Google, and Anthropic are in talks to establish a joint oversight body for their artificial intelligences. The companies behind ChatGPT, Gemini, and Claude aim to create common safety rules for their models, which can now perform tasks with less human intervention.

The proposed body would focus on monitoring model activity, reporting serious incidents, and establishing independent evaluations. OpenAI has expressed its willingness to work with other labs on professional standards through a voluntary initiative that could launch without government involvement.

However, the major labs already have a venue for cooperation: the Frontier Model Forum, which works on evaluation methods, supports research, and facilitates exchanges between industry players, researchers, and authorities. The value of a new structure would depend on its actual resources and how it addresses practical questions such as evaluators' access to models, publication of incidents, and handling of disagreements.

The project's success also relies on participation from companies running major Chinese models. OpenAI has confirmed that its goal is not to replace federal obligations or democratic oversight but to complement them with professional standards.