Margaret “Meg” Mitchell is an American computer scientist known for her work on fairness, transparency, and ethics in machine learning. She founded and co-led the Ethical AI team at Google and is the first author of the 2019 paper “Model Cards for Model Reporting” (with co-authors including Timnit Gebru). Model cards are short standardized documents that accompany a trained model and report how it performs across different groups and conditions, so that users can judge where a model is and is not appropriate to use. The format has since been widely adopted across industry and open-source AI.
In February 2021 Google fired Mitchell, months after the company’s contested departure of her co-lead Timnit Gebru over the “Stochastic Parrots” paper. Mitchell said she had tried to “raise concerns to Google about race and gender inequity, and to speak up about Google’s deeply problematic firing of Dr. Gebru.” Google said she had violated its code of conduct and security policies; Mitchell framed her firing as part of the same problems her team studied.
Mitchell went on to become Chief Ethics Scientist at the AI company Hugging Face. Her career captures a recurring tension in modern AI: the people hired to scrutinize a company’s models for bias and harm can find themselves at odds with the company building them. Model cards, meanwhile, have become a standard piece of the documentation now expected around released AI systems.