AI-related standards activities with a focus on algorithmic bias

Abstract

The last five years has seen a rapid expansion of AI-related standardisation activity, driven in part by the precautionary principle and in part by legislative attention on the potential impacts of AI. This is all despite AI not being a new technology, but more because of concerns stemming from the likely pervasive roll-out of AI technologies and questions of responsibility and accountability. This talk will provide an overview of activities at ISO and IEEE, the resources currently available and how to participate, followed up with a case study drawing on the speaker's participation in the IEEE P7003 working group on algorithmic bias considerations.