Hugging Face CEO Clément Delangue has a surprisingly low estimate for the probability of AI causing catastrophic harm.
Speaking at WIRED World Fair 2026, Delangue said his personal “P(Doom)” sits close to zero rather than the alarming figures cited by some industry figures.
He attributed most AI risk to a small number of companies developing powerful systems behind closed doors without public scrutiny.
That concentration of power, not AI itself, is what concerns him most about the technology’s future.
Delangue argued that open source AI offers a corrective by distributing development across many contributors and communities.
He said transparency allows researchers and the public to inspect models, identify flaws, and address problems before they spread.
His position places him at odds with prominent voices who warn that open models could be misused by bad actors.
Hugging Face has become a central hub for open source machine learning, hosting hundreds of thousands of shared models.
Delangue’s remarks reflect a broader debate over whether AI safety is best served by secrecy or openness.
He contends that closed development leaves oversight to a handful of firms with little external accountability.
The CEO’s near-zero doom estimate contrasts sharply with forecasts from some researchers who see serious existential threats.
His comments add weight to the argument that openness and public participation should shape how AI is built and governed.
