Rumman Chowdhury
Founder and CEO of Humane Intelligence; former head of Twitter's ML ethics team
Rumman Chowdhury built and led Twitter's Machine Learning Ethics, Transparency and Accountability team, ran the first public generative-AI red-teaming exercise at DEF CON 31, and founded Humane Intelligence to run structured public evaluations of AI systems. She served as a US Science Envoy for AI.
Key arguments & positions
- Argues models should be tested by the public who are affected by them, not only by the company that built them.
- Holds that bias bounties and structured red-teaming produce evidence that internal review does not.
- Argues accountability needs institutions and funding, not good intentions from individual teams.
Accomplishments
- Founder and CEO of Humane Intelligence, which runs structured public evaluations of AI systems.
- Built and led Twitter's Machine Learning Ethics, Transparency and Accountability (META) team.
- Organized the first public generative-AI red-teaming exercise at DEF CON 31, backed by the White House.
- Served as a United States Science Envoy for Artificial Intelligence.
Papers & key writings
Her main output is applied — red-teaming exercises, bias bounties and public evaluations — rather than a body of papers, so her positions above come from those programs and her public talks.
Links
In the library
Nothing of theirs is filed in the hub yet. Browse the full library.
Recommended next
Hand-picked from the hub based on what Rumman Chowdhury covers.
PyRIT
Microsoft's open-source toolkit for red-teaming AI systems: automated attack prompts, scoring of the responses, and repeatable runs. Free.
Why this: Covers red teaming and AI ethics too
garak
An open-source scanner that probes a language model for weaknesses — prompt injection, data leakage, jailbreaks, toxic output — and reports what it found. Free.
Why this: Covers red teaming and AI ethics too
abliteration.ai — Unrestricted AI API and Policy Gateway
An OpenAI-compatible API for unrestricted language models aimed at red teaming, security research, evaluations, and synthetic data, paired with a policy gateway for per-project keys, audit logs, and no data retention.
Why this: Covers red teaming and AI ethics too
Gray Swan Arena
Public AI red-teaming arena where anyone can try to break frontier models in timed challenges.
Why this: Covers red teaming and AI ethics too