Margaret Mitchell

Margaret Mitchell is one of the most consequential figures in AI ethics — a computer scientist who spent nearly two decades inside the industry’s biggest labs, including Microsoft Research and Google, before becoming chief ethics scientist at Hugging Face. At Google she founded and co-led the Ethical AI team alongside Timnit Gebru, pioneering tools like Model Cards that became industry standards for documenting AI systems.

Stochastic Parrots and the firing

Mitchell co-authored “On the Dangers of Stochastic Parrots” with Gebru and Emily M. Bender — the 2021 paper warning that ever-larger language models entrench bias, consume vast energy, and produce fluent text without understanding. Google forced Gebru out in December 2020 over the paper; Mitchell, who publicly defended her, was fired in February 2021. The episode became the defining scandal of corporate AI ethics, cited ever since as proof that Big Tech will not police itself.

At Hugging Face since 2021, Mitchell occupies a deliberately uneasy position: she builds ethics tooling inside an AI company while remaining one of the field’s sharpest critics. Testifying before the Senate Judiciary Subcommittee in September 2024, she described how tech companies structurally disincentivize responsible AI work and called for mandatory documentation, disclosure requirements, and whistleblower protections.

In a June 2025 Financial Times interview she dismissed artificial general intelligence as undefined hype — “vibes and snake oil” — arguing that neither “intelligence” nor “general” has a clear meaning, creating an illusion of consensus that lets companies pursue any path under the banner of progress. She warned of a “massive rift” between those profiting from AI and the workers whose creative output trains it.