“The Promise and Peril of AI”: Ajay Agrawal Speaks to The Economist
Artificial Intelligence already recommends our movies, routes our taxis, and even predicts who we might fall in love with. While it’s on track to become a mainstream technology as fundamental as computing or electricity, it still has a few challenges ahead of it. Machine-Learning Expert Ajay Agrawal stopped by The Economist’s weekly technology podcast to discuss the benefits and risks.
Ajay Agrawal is not an AI scientist himself, but an economist who specializes in artificial intelligence. This fundamental difference offers him a unique perspective, often from a higher vantage point than those working at the code level.
“The first branch is categorizing those we can call productivity-enhancing AI. They basically do what we’ve already been doing, just better, faster, cheaper,” Agrawal explains. The second set of AI takes problems that we didn’t think of prediction problems, such as driving or translation, and recasts them as prediction. But one day, AI will go beyond things we can already do and enable us to do something entirely new. While this will be an exciting development, Agrawal fears this will cause problems around the concentration of power.
“Effectively, what we are beginning to see is that AI will likely outperform humans in virtually every form of diagnosis,” Agrawal said, using the medical industry as an example. “The writing on the wall that we can see is that eventually, [...] there will be one winner. One that gets ahead of the rest, and because it’s better than the rest, more people will use it. The underlying feature of how AI works is that will lead to natural monopolies.”
To listen to the full podcast, click here.
To book speaker Ajay Agrawal for your next event, contact The Lavin Agency today, his exclusive speakers bureau.