Hugging Face CEO Clément Delangue: My P(Doom) Is 'Close to Zero'

"Most of the risk is actually created by a couple of players building behind closed doors," Delangue said, making the case for open source AI.

In an interview with WIRED (published November 4, 2026), Hugging Face CEO Clément Delangue pushed back on the existential-risk framing that has dominated AI safety debates, arguing that his personal probability of doom—commonly abbreviated as P(Doom)—is "close to zero."

The Core Argument

Delangue's position hinges on a simple inversion of the conventional risk narrative. Rather than viewing open source AI as a hazard, he frames closed development as the primary source of danger. "Most of the risk is actually created by a couple of players building behind closed doors," he said, underscoring the need for transparency and broad access to AI systems.

For Delangue, open source is not merely a distribution strategy but a safety mechanism: distributed scrutiny, wider participation, and the absence of a small number of actors making unilateral decisions about powerful models.

Why It Matters in 2026

The comments land amid an intensifying debate over open-weight models. In 2026, regulators in the US and EU continue to grapple with how to treat openly released model weights under emerging AI governance frameworks, while frontier labs remain divided on how much openness is prudent. Hugging Face, as the central hub for open models and datasets, has a direct stake in that outcome—and Delangue has become one of the most prominent voices arguing that openness and safety are compatible, not opposed.

The Takeaway

Delangue's near-zero P(Doom) stands in sharp contrast to the more alarmist estimates offered by some frontier-lab leaders and safety researchers. His argument reframes the debate: the question is not simply whether AI is dangerous, but who gets to build it, who can inspect it, and who bears the risk when development happens out of public view.

Reporting by Will Knight for WIRED.

via Wired AI