Timnit Gebru is an AI researcher best known for showing, with hard numbers, that AI systems can work great for some people and badly for others. She’s also a co-author of the paper that gave us the phrase Stochastic Parrot, the idea that language models can sound smart without understanding anything.
She’s one of the sharpest critics of how big tech companies build AI. That’s made her a lot of fans and a few very powerful enemies.
Who they are
Gebru was born in Addis Ababa, Ethiopia, came to the U.S. as a teenager and earned her PhD at Stanford in 2017, studying under Fei-Fei Li. She worked at Microsoft Research and then co-led Google’s ethical AI team. In 2017 she co-founded Black in AI, a group for Black researchers that started as a workshop at a major AI conference. She was also lead author on “Datasheets for Datasets,” a widely used proposal for documenting what’s inside training data.
Today she runs DAIR, the Distributed AI Research Institute, which she founded in December 2021 with $3.7 million from foundations including MacArthur and Ford. In September 2026 she received the Right Livelihood Award, sometimes called the “alternative Nobel,” for “challenging concentrated power in artificial intelligence.”
Why they matter
She put numbers on AI bias. Her 2018 study with Joy Buolamwini, called Gender Shades, tested three commercial face recognition systems. Error rates hit 34.7% for darker-skinned women versus 0.8% for lighter-skinned men. Several companies changed their products after it came out.
Her exit from Google became a big story. In December 2020, Google said it accepted her resignation. Gebru says she never resigned and was fired over a paper about the risks of large language models. That paper, “On the Dangers of Stochastic Parrots,” came out a few months later in 2021 and has been cited constantly since.
She works outside big tech on purpose. DAIR studies how AI affects the people it’s used on, not just how well it scores on tests. She has a memoir, Deep Unlearning, coming in February 2027.
What they’ve said or done
In an October 2026 interview with Democracy Now!, Gebru argued that a lot of AI policy is being shaped by the companies it’s supposed to regulate. “What is actually happening is regulatory capture,” she said.
You don’t have to agree with her to see why she keeps getting invited on. She’s been early on this stuff more than once.
