Machine learning offers great potential to make predictions that can be better than those of humans. But it can also lead to outcomes just as bad or worse, since training datasets overwhelmingly reflect unfair human biases and practices. In this session we will discuss bias and unfairness in AI, and what can be done about it.
What are the ways that machine learning embeds human bias? What are the implications at scale? What can be done to negate or reverse algorithmic harm?