Fair AI: Supporting Social Justice and Equality by Eliminating Bias in AI

As AI-based systems make more decisions for us, there is an important need to ensure that these systems do not reproduce and exacerbate existing social biases and forms of discrimination.

The Fair AI team proposes to develop a platform that will evaluate and identify evidence of bias in relation to gender, race and political ideology in AI training sets. There are growing calls from both governmental institutions and industry for solutions to mitigate algorithmic bias. The Fair AI team will develop a scalable system that could be employed across a range of applications to evaluate training data for evidence of bias, thus working towards the goal of fairness in AI.