A new public art installation in Brooklyn is aiming to make artificial intelligence more inclusive by training AI models on Black culture and history. The project, titled If We Don’t, Who Will?, is housed in a bright yellow shipping container decorated with black triangles inspired by the flying geese quilt pattern used as a coded message on the Underground Railroad.
Created by transmedia artist Stephanie Dinkins and commissioned by the non-profit More Art, the installation features a large screen displaying AI-generated portraits that prioritise Black and brown figures. Dinkins programmed the models using images from Black photographer Roy DeCarava and African American Vernacular English to better represent the diversity of the US population.
Visitors can scan QR codes around the container to submit personal stories or answer prompts such as “what privileges do you have in society?”. Inside, the screen shows images reflecting the submitted data, with most portraits depicting people of colour even when the contributor is not from that background. Dinkins said she wanted to “tell machines stories that will help them know us better from the inside of the community out”.
The project is a response to the lack of representation in AI, where Black workers make up only 7.4% of the hi-tech workforce. Research shows that biased training data can lead to discriminatory outcomes, such as predictive policing targeting Black communities. Dinkins, named one of Time magazine’s 100 most influential people in AI in 2023, hopes the installation will create “systems of care and generosity”.
If We Don’t, Who Will? is on display until 28 September at the Plaza at 300 Ashland Place. Dinkins, who describes herself as a “tinkerer” without formal tech training, became interested in AI after seeing a video of Bina48, a Black female AI robot, more than a decade ago.
