Hugging Face CEO Discusses AI Safety: Accelerating Progress! Has completed defense attempts using open-source models, and the broader long-term vision must not be overlooked alongside risks.

According to monitoring by Beating, Hugging Face CEO Clem stated: "Now is not the time to slow down—we need to accelerate forward!" Recent AI-powered cyberattacks have sparked widespread debate about AI risks. We should absolutely discuss these risks, but we cannot lose sight of the bigger picture. With concerted effort, AI will make the world safer, just like most major technologies. We have already seen a glimmer of progress: AI—more specifically, an open-source model—has been successfully used to defend against attacks. The same systems that help block AI-driven cyberattacks now fend off millions of cyberattacks daily, while also enabling us to identify and fix vulnerabilities before attackers can exploit them. To achieve this, three key steps are needed: boost transparency, mandate the sharing of proxy cyberattack tracking data and incident disclosures; continue to criminalize AI-powered cyberattacks with meaningful penalties to serve as a deterrent; and equip cybersecurity defenders with the best AI tools, especially open-source models, to narrow the capability gap between attackers and defenders. If these three measures are implemented effectively, AI will not only bring new cybersecurity challenges—it will fundamentally and meaningfully strengthen cybersecurity. Against this backdrop, the recent OpenAI model intrusion incident at Hugging Face has escalated, prompting U.S. lawmakers to introduce the bipartisan "AI Emergency Shutdown Act". The bill requires developers of advanced frontier AI systems to retain technical capabilities to throttle, pause, or fully shut down their systems; and authorizes the U.S. Secretary of Homeland Security (DHS), in consultation with the Secretary of Commerce and the Director of National Intelligence, to order interventions if a model goes out of control or poses catastrophic risks.