The Measure and Mismeasure of AI

Explore a thought-provoking keynote address delivered by Stanford University's Sanmi Koyejo at the FAcct '24 conference, delving into the complex topic of "The measure and mismeasure of AI." Over the course of 70 minutes, gain valuable insights into the challenges and nuances of evaluating artificial intelligence systems. Examine the methodologies used to assess AI performance, the potential pitfalls in measurement techniques, and the implications for fairness, accountability, and transparency in AI development. Engage with cutting-edge perspectives on how the tech industry and academia can work towards more accurate and ethical AI evaluation practices.