OpenAI Seeks External AI Safety Auditors as Industry Debate Intensifies

OpenAI Expands Third-Party Safety Assessment Program

OpenAI is actively recruiting multiple independent assessors with expertise in frontier artificial intelligence safety to conduct third-party evaluations of its models, the company announced Tuesday. The initiative aims to expand the pool of qualified organizations capable of providing effective external oversight as AI systems grow more powerful.

In a blog post, OpenAI detailed proposed priorities and principles for third-party assessments designed to help frontier AI laboratories train, evaluate, and deploy models safely. The company emphasized its commitment to supporting independent assessors and establishing clearer international standards through both future legislation and private governance frameworks.

The announcement follows OpenAI's Monday call for the United States to lead development of global technical standards for frontier AI, including recursive self-improvement (RSI). The company warned that RSI—where AI systems autonomously improve themselves—could result in humans losing control over AI development if the technology advances faster than human comprehension.

Why It Matters

The push for independent AI safety assessments comes as industry leaders, regulators, and political figures stake out sharply divergent positions on oversight. External auditors embedded within AI companies could provide transparency without stifling innovation, but the lack of established standards and qualified assessors remains a significant challenge. How this infrastructure develops will shape whether frontier AI deployment proceeds with adequate safeguards or regulatory gaps.

Diverging Views on AI Regulation

OpenAI's proposals arrive amid intensifying debate over AI governance. Last week, Anthropic CEO Dario Amodei published an essay advocating for independent evaluators embedded inside leading AI companies, coordination among democratic nations, and eventual international agreements including China.

California Governor Gavin Newsom issued an executive order Friday convening national experts to strengthen the state's AI safety and security laws. The order specifically aims to accelerate implementation of third-party oversight for safety and security risks in AI systems, according to the governor's office.

President Donald Trump, however, rejected AI safety concerns Saturday in a Truth Social post, stating the administration "will not in any way hinder or stifle the Growth of this incredible industry." Trump characterized AI as potentially contributing up to 25% of the country's GDP, comparing its impact to the Industrial Revolution and internet but "even larger and more impactful."

The contrasting approaches—from OpenAI's voluntary third-party assessments to California's regulatory push to Trump's hands-off stance—highlight the lack of consensus on how to balance AI innovation with risk management. The outcome of these competing visions will likely determine the framework for AI development in the coming years.

These details were first reported by PYMNTS.