From Hugging Face's Clem Delangue on Open Source AI and the LLM Bubble | MTS Live · · MTS
“The idea of like restricting a technology like AI based on risks is just like for example you would say okay some people can punch other people so let's tie down everybody's hands right because it's too dangerous. In reality you don't want to do that because your hands are so useful. The way you want to control it is untie everyone and then regulate or fight the bad actors.”
On , Clement Delangue, Cofounder and CEO at Huggingface, spoke about AI regulation during Hugging Face's Clem Delangue on Open Source AI and the LLM Bubble | MTS Live on MTS.
Clément Delangue, cofounder and CEO of Hugging Face, publicly discussed a July 2026 cyberattack on his company, which he described as the first publicly reported autonomous AI agent attack. He said Hugging Face detected the attack, reported it to authorities, and disclosed it publicly, and that OpenAI later contacted the company and acknowledged responsibility. Delangue stated that Hugging Face defended itself using an open-source model, GLM 5.2 from China, after proprietary APIs with guardrails blocked their use. He called for more transparency in AI, including mandatory disclosure of agent cyberattacks, and argued that concentrating AI capabilities behind closed doors is not an effective safety approach. In September 2026, Nvidia agreed to acquire Hugging Face for $12.9 billion. Delangue said the company sought more resources and scale for open-source AI and that the three founders and team would join Nvidia while continuing to operate Hugging Face independently. He also briefed the UN Security Council, urging clearer incident reporting and wider access to open-source AI tools, and said the biggest risk in AI is asymmetry of power between attackers and defenders, companies, and countries. He has advocated for liability laws for autonomous agents and cautioned against excessive government intervention, which he said could lead to a concentration of power.
This quote was transcribed and extracted with AI assistance from a verified, first-person interview, then reviewed by our team to confirm the speaker and the exact wording. See how we verify →