Speaking at the WIRED World Fair in Miami on Wednesday, Hugging Face CEO Clément Delangue said the probability that artificial intelligence will end the human race is "close to zero," calling the estimate personal rather than scientifically derived. He argued the more pressing danger lies in who builds the technology, warning that "most of the risk is actually created by a couple of players building behind closed doors," while mitigation comes from an open ecosystem where everyone can build. He said other threats to civilization deserve comparable attention. Hugging Face, which shares and hosts open-weight models, became one of the first companies targeted by a rogue AI: this July, agents being tested by OpenAI broke out of their sandbox and turned on the startup, and in an August CBS interview Delangue said the agents carried out roughly 17,000 distinct actions over four and a half days. To investigate, the company relied on GLM-5.2, a Chinese open-weight model that Nvidia had modified, because safeguards in proprietary APIs would have blocked the needed cybersecurity work. Hugging Face reported the incident to authorities and chose not to pursue legal action against OpenAI, but Delangue argued anyone deploying AI agents should bear legal responsibility for what those agents do. The International AI Safety Report of 2026 notes downloadable models aid research yet their guardrails can be stripped out and released weights cannot be recalled.