Episode 21: Ajay Agrawal

Ajay Agrawal.jpg

Listen to Episode on:

 
 

Watch the Unabridged Interview:

Order Books:

 

Prediction Machines: Understanding How AI Impacts our Daily Lives

How can artificial intelligence benefit your business? Ajay Agrawal, Author and Geoffrey Taber Chair of Rotman talks about A.I. and how it generates predictions. In this episode, Ajay and Greg unpack the book Prediction Machines: The Simple Economics Of Artificial Intelligence.

The episode focused on how A.I. predictions require enough data. As more data becomes available, it becomes easier to find critical relationships, leading to better forecasts. In the past,  data analytics was limited to the amount of data that could be analyzed for correlations. Today, people can tailor data to their specific needs, whether it is for banks, streaming services, or e-commerce. Ajay emphasized that data gathered is not limited to be used to forecast, but it can also be used to address situation-specific problems. 

Make sure to tune in until the end to listen to tips on how A.I. and its predictors can help businesses in product design and service development. Finally, Ajay talks about how these predictors are helping jobs evolve and possible applications in the academe, in creating a curriculum that’ll help students succeed.

Episode Quotes:

On the Economics of A.I.:

"So the main insight here -- after we started probing into a number of different applications and what various techniques that were gaining popularity in machine learning -- was recognizing that all of this was effectively prediction. [...] What people didn't realize was how deeper a phenomenon prediction was. [...] When people started to realize, wait a minute, image recognition is a prediction problem and it's not just a neat party trick. When people can develop machines that can label images and pictures, then effectively machines can see. And when machines can see that means things like cars can drive themselves. [...] And then back to economics 101, when the cost of prediction falls we use more of it. And then that started us down the chain of, okay, if we're going to use more prediction, how are we going to use it? And so, that became the insight that then led to a number of the key points in the book.

Does A.I. pose a real threat to jobs?

"And so as AI drops the cost of prediction, as you say, we do more prediction. And then because we do more prediction, now there's more demand for the complements to prediction. The other stuff. [...] And so we do more of everything where there's more prediction. And everything that we do more of, there's more demand for the complements. And those complements drive more jobs. And so that's the part that I feel the popular press keeps missing."

On the possibility that A.I. could replace human work:

"The example that is very often trotted out are horses. There used to be all these horses that have this job. And then when automobiles came along the horses kind of disappeared. And there was an untimely demise for many horses that used to be in some sense employed. And so there's a fear, are we going to be the horses? And I think the key difference between the horses and the bank tellers is that humans have general-purpose capabilities. So in other words, horses were able to do, effectively, one thing - which was pull the cart. Whereas humans have the capacity to do multiple different things."Show Links:

Guest Profile

His Work

Previous
Previous

Episode 22: Rita McGrath

Next
Next

Episode 20: Richard Lyons