Clement Delangue is meeting with Washington policymakers on Sept. 16 to discuss a new class of security risk. The Hugging Face chief executive officer identified the company as the first publicly disclosed victim of a breach carried out by an AI agent and will speak at Politico's Decoded summit to share details on the threat.

The CEO's proposal for government officials includes increasing transparency and maintaining the illegality of AI powered attacks through meaningful penalties. Delangue argues that defenders must be equipped with the best AI tools to combat automated attacks to mitigate the risks associated with agent based vulnerabilities.