Cognitive scientist, Prof. Abebe Birhane, joined iAfrikan for an episode of iAfrikan Bytes, live at the University of Ghana during Deep Learning Indaba.
Awarded Time Magazine's "100 Most Influential People in AI" Abebe helped establish the field of auditing publicly available AI training datasets. This is very important because it keeps dangerous datasets from being gathered, which could cause AI to inherently be biased, sexist, or racist.
In our conversation, Abebe shares insights on her recent paper, on how she and co authors came to the startling conclusion that AI models trained on larger datasets are more likely to display harmful biases and stereotypes.
In addition, she discusses the importance of the Kambule Doctoral Award, Alele-Williams Masters Award, and Maathai Impact Award on the fringes of the Indaba in her role as an awards committee member.
