Developing AI apps free from bias crucial to avoid analytics errors

D&S Researcher Madeleine Clare Elish discusses the implications of biased AI in different contexts.

She said when AI is applied to areas like targeted marketing or customer service, this kind of bias is essentially an inconvenience. Models won’t deliver good results, but at the end of the day, no one gets hurt.

The second type of bias, though, can be more impactful to people. Elish talked about how AI is increasingly seeping into areas like insurance, credit scoring and criminal justice. Here, biases, whether they result from unrepresentative data samples or from unconscious partialities of developers, can have much more severe effects.