“AI is a lie”

In this interview from O’Reilly Foo Camp 2019, Eric Jonas, assistant professor at the University of Chicago, pierces the hype around artificial intelligence.

Highlights from the interview include:

Jonas argues that “AI is a lie”—meaning that our expectations far outsize the reality of what’s currently possible. One of the issues arising from that disconnect is a level of corporate investment in the research process that hasn’t been seen before. He argues this has led to “a lot questions about whether robots are going to take our jobs, and these sorts of things, which are all quite premature.” (00:41)

Questions of ethics and what role it should play are increasingly arising in machine learning and AI research, especially in the area of science applications. While Jonas applauds such inquiry and thinking deeply about the social ramifications of AI research, he is concerned the questions might be reinventing the wheel: “The data collection itself often has serious ramifications that we’ve all been wrestling with for 15 years. It’s hard to figure out, necessarily, when a linear model might be biased, or what exactly that means—there’s been a tremendous amount of excitement around these questions. But I’m trying to also remind everyone that the challenges there have existed for a while, and there’s a lot of good research and scholarship on addressing them.” (2:12)