Responsible Innovation: Implementing Google’s AI Principles through Adversarial Testing
This session will introduce Google’s AI Principles, which drive the development of responsible AI throughout Google. We will provide some examples of how we spot issues relevant to the principles. Then we will lead the participants in a workshop focused on one of the ways we put the principles into practice, by conducting adversarial testing on products. In this testing we take on the user perspective to identify potential fairness issues and advise teams on mitigations and other steps to improve product fairness.