The BBC has developed AI principles and a risk management framework to guide the corporation’s approach to the rapidly changing AI landscape. But turning key principles into practice means diving into the challenging world of evaluating the technical performance of AI. In this talk James and Oli will share the BBC’s emerging approach to AI evaluation, with practical examples of testing AI services and measuring public service values like representation and universality in personalised platforms such as iPlayer. 

Ever go to conferences hoping for practical insights that can actually help you do responsible AI better day to day, but come away empty handed after yet another high-level discussion about the EU AI Act? Well this session is for you. It will be interactive, it will be fun, and most of all it will be a chance to get answers to your burning questions about how to survive and thrive as a responsible AI practitioner doing it for real in the fast-changing, complex world of AI. 

Lead by the BBC's Responsible AI team