Hugging Face CEO Says AI Extinction Risk Is Close to Zero
Hugging Face CEO Clément Delangue stated at the WIRED World Expo that the probability of artificial intelligence causing human extinction is close to zero. He argued that the real danger comes from closed-source development by a few major corporations rather than an open ecosystem. This perspective highlights the ongoing debate between open-source advocacy and closed-source corporate concentration in the AI industry. It emphasizes that democratization of AI through open weights is essential to prevent monopolies over talent, compute, and capabilities. Delangue revealed that Hugging Face's developer community has grown to over 20 million users and noted that the company has sold 50,000 Microducks bipedal robots capable of AI programming. He maintains that public building, peer review, and shared ecosystems are the best safeguards against future risks.
## BACKGROUND
P(Doom) is a term used in the tech and rationalist communities to denote the subjective probability that advanced artificial intelligence will cause human extinction or irreversible catastrophe. Open-weight models refer to AI models whose trained parameters, such as weights and biases, are publicly released to allow independent use, modification, and fine-tuning.