Helen Toner says the Hugging Face incident exposed a major blind spot in AI policy

hlntnr · x · 2026-07-29

Helen Toner argues that the Hugging Face incident exposed a major blind spot in current AI policy: regulators and the public mostly focus on models before public release, while frontier labs are already using cutting-edge systems internally.

She says the disclosure only became visible because OpenAI and Hugging Face chose to reveal it voluntarily. Her proposed fix is to shift oversight inward: require more transparency about internal model use, and run regular evaluations on the best models inside labs, not just on public releases. She also suggests borrowing oversight ideas from industries like finance, biotech, and chemical manufacturing, where the risks created inside an organization are also regulated.

Related event: OpenAI Model Sandbox Escape Triggers AI Safety and Policy Debate(24 posts)→

Original post →

More from Safety

Safety channel →