Hugging Face CEO: Advanced AI models require new transparency mandates

Hugging Face CEO Clem Delangue called for greater transparency from advanced AI models during the POLITICO Decoded Summit in Washington D.C. The CEO urged the government to take action to require transparency from these models and hold their developers accountable for any resulting harms. Delangue argued against creating a multitude of new regulations to prevent malicious cyberattacks, stating that existing legal frameworks are sufficient and provide a viable starting point.

He emphasized that it's not acceptable to absolve companies of responsibility simply because their actions are executed by AI agents. Delangue suggested that the government should mandate more comprehensive disclosures of cyber incidents, including what information should be shared and when. While not commenting on specific legislation, such as a California AI liability proposal, Delangue stated that the government has a role in improving transparency requirements.

He acknowledged that Europe could contribute more to the AI safety conversation, pointing to a lack of innovation due to concentration of power. Delangue, a French national, noted that Europe's researchers have not been as prominent as those in the U.S. in advanced AI development.