A new paper from the MIT Media Lab’s Joy Buolamwini shows that three commercial facial-analysis programs demonstrate gender and skin-type biases, and suggests a new, more accurate method for evaluating the performance of such machine-learning systems.

Companies pledged money and support for people of color. But some say they still face a hostile work environment for speaking out or simply doing their jobs.

The use of artificial intelligence (AI) presents risks to equality, potentially embedding bias and discrimination. Auditing tools are often promised as a solution. However our new research, which examines tools for auditing AI used in recruitment, finds these tools are often inadequate in ensuring compliance with UK Equality Law, good governance and best practice. We argue in this report that a more comprehensive approach than technical auditing is needed to safeguard equality in the use of AI for hiring, which shapes access to work. Here, we present first steps which could be taken to achieve this. We also publish a prototype AI Equality Impact Assessment which we plan to develop and pilot.