Wonderfest: Human Values in A.I.

How do we ensure that artificial intelligence (AI) reflects and enacts human values? UC Berkeley’s Dylan Hadfield-Menell will explore how AI is applied in the world, and how it can go wrong - even terribly wrong. He’ll examine potential solutions that draw on an analogy with misaligned incentives in economics: in particular, the “gaming” of performance measures like standardized tests scores. He will also show how the future of AI is related to the Wells Fargo account scandal … and even to “fake news.”

Speaker: Dylan Hadfield-Menell, UC Berkeley