Rumman Chowdhury

Rumman Chowdhury is a data scientist with a PhD in political science who became one of the most influential figures in responsible AI. She founded the Responsible AI practice at Accenture, then led Twitter’s META team (Machine Learning Ethics, Transparency, and Accountability) until the 2022 Musk takeover — running, among other things, an “algorithmic bias bounty” that let outsiders probe the platform’s image-cropping model.

Building the brakes

In 2023 Chowdhury co-founded Humane Intelligence, a nonprofit that runs large-scale red-teaming and bias-bounty exercises — including the landmark generative-AI red-teaming event at DEF CON, backed by the White House. In 2024 she was named the first United States Science Envoy for Artificial Intelligence, carrying AI-governance engagement to emerging markets. Her signature argument, delivered in congressional and state testimony, is that good rules accelerate rather than hinder adoption: “brakes help you drive faster.”

Chowdhury is unsparing about industry practice — she dismisses one-off “checkbox” bias scans, warns that AI already mediates hiring, housing, and healthcare “often with little transparency,” and demands enforceable audits over “voluntary ethics pledges.” Yet her actions build the machinery of trustworthy AI rather than oppose the technology: evaluation ecosystems, whistleblower protections, and government capacity. Hers is a pro-accountability optimism — the conviction that AI is worth getting right.