What if the data used to train AI-powered systems were stewarded with care and reflected the fullness of the human experience?

Dominant digital platforms increasingly shape what is “relevant,” “true” or accepted as common knowledge.

Scroll to learn

The Digital Landscape

There are 5.56 billion internet users worldwide, representing 67.9 percent of the global population.

Source: Digital 2025: Global Overview Report, DataReportal (2025)

0.0%

Wikipedia is one of the largest single sources of training data for AI-powered systems.

Source: “Wikipedia’s Moment of Truth”, NYT Magazine (2023)

0%

of the content on Wikipedia is crafted by male contributors who reside in the United States (20%), Germany (12%), and Russia (7%).

Source: “Who Writes Wikipedia?,” Wikipedia (n.d.)

Algorithmic Bias

The source of knowledge in the world is filtered through the lens of a small and very specific group of people.

The lack of expansive training data in AI systems puts under-considered populations at risk of experiencing algorithmic bias, digital erasure and misrepresentation within default narratives. A clear example is Alexa's definition of Black women.

press play to Watch the video below

Our Intervention

Good Mirrors Index is a digital encyclopedia and archive of Black femme consciousness.

The Index centers biographies, articles, and stories about Black women to expand the data used to train AI-powered systems.

The Index will inform the creation of a dataset* that preserves Black women's wisdom, aesthetics, and more.

*Good Mirrors Index is a precursor to Good Mirrors AI, a large language model built from Black women's knowledge.

We are building a dedicated community of stewards around this project:

Community Griots
Librarians
Archivists
Historians
Futurists
Poets
Storytellers
Artists
Community Builders and Leaders
Students
Photographers
Black Feminist Scholars
Cultural Provocateurs
AI Enthusiasts and Experts
Engineers
Funders
...and YOU.

This platform will:

Inform the build of an AI-tool designed to preserve Black women's shared knowledge, wisdom, aesthetics, and more-ensuring that Black women's legacy is accessible to future generations.

Impact data governance and algorithmic construction for discoverability of more "human" and under-told narratives of Black women online and expand the data that is used to train AI-powered systems on how to depict Black women and girls on the internet.

Help to combat stigmas that negatively impact how Black women are perceived and treated in society. Stigmas that contribute to chronic stress and anxiety faced by over 23% percent of the 21 million Black women in the United States.

Join Us

Steward the Internet with Care

Thank you! Your submission has been received!

Oops! Something went wrong while submitting the form.