# Treasury Secretary Says OpenAI Managers Are to Blame for Hugging Face Breach, Not AI Agents URL: https://technosports.co.in/treasury-secretary-says-openai-hugging/ Published: 2026-09-22 Updated: 2026-09-22 Author: Reetam Bodhak The Treasury Secretary says OpenAI’s managers are to blame for the Hugging Face security breach — not autonomous AI agents. Those official remarks came Tuesday, September 22, 2026. The statement puts responsibility on **human leadership at [OpenAI](https://technosports.co.in/tag/openai/)**, not the models the company builds. Whether a sitting U.S. Treasury Secretary has ever publicly pinned a major AI security failure on a corporate management layer before? Not clear from the available record. ![](https://technosports.co.in/wp-content/uploads/2026/09/Treasury-2.jpg) ## What the Treasury Secretary Says About Who Owns the Failure The Treasury Secretary cast the incident as a governance breakdown, not a technical one. That distinction carries real weight. If autonomous agents were at fault, regulators would lean toward model-level restrictions. Blaming managers points somewhere else instead: deployment oversight, access controls, internal review processes. On the available record, the remarks didn’t say which controls failed or [how](https://technosports.co.in/improve-visibility-across-your-enterprise/) far the exposure reached. Technical specifics of the breach remain unconfirmed. The platform affected was Hugging Face. It’s an open model hub and community hosting hundreds of thousands of public model repositories, so the incident touched infrastructure thousands of independent developers rely on daily — not one closed corporate system. **Verdict:** Blaming managers, not models, shifts the regulatory target from AI architecture to corporate accountability. ## Why Hugging Face Changes the Stakes **Hugging Face** is no fringe player. It’s the default distribution channel for open-weight models, which turns any compromise there into a supply-chain problem rather than a single-company event. That’s exactly why the [Base Hugging Face safety debate](https://technosports.co.in/base-labs-open-weight-safety/) has run for months without a clean resolution. And the competitive backdrop sharpens the pressure. OpenAI’s rivals have spent 2026 arguing that safety failures are organisational, not architectural — a theme running through coverage of [Anthropic's reported GPT-6 counter](https://technosports.co.in/anthropic-new-claude-model-counter-gpt-6-astra/). Assigning blame to managers supports that reading. It also hands OpenAI’s critics a government-sourced talking point. Worth noting: the remarks land while Washington debates how much liability should sit with executives when AI systems misbehave. Our earlier reporting on the [lawsuit alleging an AI slowdown agreement](https://technosports.co.in/ai-slowdown-lawsuit-anthropic-openai/) raised the same structural question about who answers for deployment decisions. OpenAI publishes its own security and policy positions through its [official blog](https://openai.com/blog). The company hasn’t responded publicly to the Treasury Secretary’s characterisation.