In this week’s Creative AI podcast Jon and Nina talk with former neuroscientist and now design PhD researcher, Hannah Korsmeyer to discuss issues of bias in AI. Using Hannah’s research intervention “Gendertron” they examine how playful AI devices can be used to start conversations around issues and assumptions about gender and other biases in AI and technology.
We profiled Hannah’s work in an earlier ai.sensilab blog post where she told us about an earler version of her Gendertron project. We discuss some of the ideas behind implicit bias and try to understand how AI researchers can better incorporate awareness of bias in the systems they develop. Can AI ever be free from bias? Listen to hear the team’s ideas on bias in AI.
If you’re interested in this area, here are some related papers and articles on language:
- Gender, Genre, and Writing Style in Formal Written Texts
- Neural networks learning language’s hidden bias
- Joanna Bryson’s blog on bias from language corpora
Please let us know your opinions about Bias in AI via twitter and we also welcome suggestions for topics you’d like to hear about in future episodes.