OpenAI Recruits External Watchdogs to Police AI Risk

OpenAI said this while outlining its proposed priorities and principles for effective third-party assessments to help all frontier AI labs train, evaluate and deploy models safely.

“We are committed to supporting independent assessors and establishing clearer, shared international standards — both through future laws and private governance initiatives — for effective third-party assessments,” OpenAI said in the post.

OpenAI’s publication of these proposals came a day after the company said that the United States should lead an international effort to develop global technical standards for frontier AI, including recursive self-improvement (RSI).

While announcing this suggestion in an article published Monday (Sept. 21), OpenAI said that RSI could result in humans losing control over AI development if this fully autonomous development of AI outpaces humans’ ability to understand it. The company said it’s necessary to decide now whether and how to proceed with the technology.

OpenAI’s latest proposals also came about a week after Anthropic CEO and Co-Founder Dario Amodei published an essay that proposed an AI safety plan that would include independent evaluators embedded inside leading AI companies, coordination among companies in democratic countries and eventually an international agreement that includes China.