Blog

This is the blog section of Professor Ajay Agrawal’s official website. Here you will find articles featuring interviews with Professor Agrawal about the economics of artificial intelligence, as well as news updates about his work and field of studies.

Professor Agrawal Nominated for Thinkers50 Digital Thinking Award

Professor Ajay Agrawal along with Professors Joshua Gans and Avi Goldfarb have been nominated for the Digital Thinking Award. The award is part of the Distinguished Achievement Awards by Thinkers50, the global leaders in ranking management ideas.

The three coauthors of the bestselling ‘Prediction Machines: The Simple Economics of Artificial Intelligence’ (Harvard Business Review Press, 2018) debunk some of the mystery surrounding AI and show how its transformative potential derives from the technology's ability to make better and more accurate predictions .

Other Rotman School of Management professors to be nominated include András Tilscik and his colleague Chris Clearfield for their book ‘Meltdown: Why Our Systems Fail and What We Can Do About It’ (Penguin Press, 2018), which won the 2019 National Business Book Award in Canada.

According to its website, ‘Thinkers50 is the world’s most reliable resource for identifying, ranking, and sharing the leading management ideas of our age. We are based in London, with partners and affiliations around the globe. Our ambition is to provide innovative access to ideas with the power to make the world a better place.’

Read more at the original Rotman article or download the official press release with the full list of nominations here.

Thinkers50 - Distinguished Achievement Awards.jpg

'Prediction Machines' Nominated for George R. Terry Book Award 2019

Prediction Machines: The Simple Economics of Artificial Intelligence” has been selected as a finalist for the 2019 George R. Terry Book Award.

This award is granted annually to the book judged to have made the most outstanding contribution to the global advancement of management knowledge during the last two years. This year’s committee membership included Ann Langley, Michael G. Pratt, Paul Leonardi, Sharon A. Alvarez and Yves Doz, from HEC Montréal, Boston College, University of California at Santa Barbara, University of Pittsburgh and INSEAD, respectively.

Prediction Machines is one of five book finalists being recognized this year from among a field of over fifty books nominated from publishers around the globe.

This is the message Professor Agrawal received from the committee:

“Your book stood out to the committee as exemplary because we found it to be a remarkable synthesis of knowledge on artificial intelligence from a decision-making perspective, a topic that is important and extremely timely. The way it frames the role of AI is very useful and backed by good research.”

The Academy of Management acknowledges all award finalists and winners through multiple media channels, including a recognition video that is displayed during the Annual Meeting, acknowledgement on the AOM website, publication in the AcadeMY News and in the 2019 Annual Report.

The award presentation to recognize this prestigious accomplishment will take place at the Academy’s Annual Meeting in Boston, Massachusetts, USA on August 11, 2019.

More information on the George R. Terry Book Award can be found on the Academy of Management website here.

Prediction Machines Book Released in Japan

Following on from our January update regarding the coverage of Prediction Machines and Professor Agrawal in India, this article highlights more international exposure and acclaim, this time from Japan.

‘Prediction Machines’ was released in Japan in early February 2019 with a translated title of ‘A Century of Predictive Machines - A New Economy Driven by AI’ ( ‘予測マシンの世紀―ーAIが駆動する新たな経済’ )

Professor Ajay Agrawal (アジェイ アグラワル) appeared in two recent articles for project.nikkeibp.co.jp. Test out your Japanese language skills by following the links below:

  1. In the past year, AI began to change business - "Century of Prediction Machine" Author's Appeal, AI's Impact (Part 1)

  2. In the future when Japan can play an active part in AI - "Century of Prediction Machine" Author's Appeal, AI's Impact (Part 2)

May 2019 Update

Professor Agrawal has been featured in three more articles in Japanese, including Forbes Japan:

  1. Article in Nishi Nippon

  2. Article in Forbes Japan

  3. Article in Beauty Tech Japan

Click here to view a rare 6-page special in Nikkei Business on Professor Agrawal and AI.

Professor Ajay Agrawal - Image from the Project Nikkei Article

Professor Ajay Agrawal - Image from the Project Nikkei Article

Prediction Machines Coverage in the Print Media in India

The book ‘Prediction Machines - The Simple Economics of Artificial Intelligence’, co-written by University of Toronto professors Ajay Agrawal, Joshua Gans, and Avi Goldfarb, was released in April 2018. It quickly achieved success as it became an Amazon Canada best-seller in the non-fiction category. It was also featured on numerous publications in the US such as the New York Times, the Economist, Wall Street Journal, and the Financial Times.

What is refreshing about the book, and added to the debate concerning the impact artificial intelligence will have on business and society, were a number of factors:

  1. This is a book written by economists about artificial intelligence and so gives a different perspective to a field dominated in the media by the thoughts of theoretical physicists, technologists, and business leaders. Economists have a habit of looking at what a new technology will make cheaper and the ramifications this will have on the economy and wider society.

  2. The book is also a look at AI in its current state and what it means for business today. It is less concerned with general AI and all the huge philosophical and moral issues surrounding that topic.

  3. Finally, the book gives business leaders a framework into how to use AI to adapt or even revolutionize their current business models. It is a practical book that gives access to strategical insights and tips to implement and use today or in the near future.

As well as getting a lot of press coverage in North America, because of the reasons outlined above, the book achieved widespread global coverage, including being featured in numerous broadsheets in the print media in India.

This article provides six examples of this:

1.‘India Could be the Data Hub for the New AI Arrival’ by Anirudh Bhattacharyya - Hindustan Times, New Delhi - April 28, 2018

2. ‘As AI Takes Over’ - Indian Management Monthly, Mumbai - May 2018

3. Prediction Machines Book Review - The Hindu, Chennai - June 6, 2018

4. ‘From Agriculture to Art: The AI Wave Sweeps In’ by Steve Lohr - Deccan Herald, Bengaluru - October 23, 2018 (originally featured in the New York Times)

5. ‘Robots Won’t Take Away Everyone’s Jobs’ by Noah Smith - The Financial Express, New Delhi - November 5, 2018 (originally featured on Bloomberg)

6. ‘Mind and Machine’ - Business Standard, New Delhi - November 10, 2018

Prediction Machines, Business Standard, New Delhi, Nov1018.jpg

Why Predicting the Future is About to Become Cheaper

In 2018, Professor Agrawal appeared on Big Think's awesome series of videos where experts are asked to answer a pertinent question or speak on a topical theme related to their field or industry.

In this video, Agrawal explains why predicting the future is about to become cheaper and what it means for companies in the era of artificial intelligence.

Below the video, you will find the transcript of Professor Agrawal’s interview.

“I think economics has something to contribute in terms of our understanding of artificial intelligence because it gives us a different view.

So, for example if you ask a technologist to tell you about the rise of semi-conductors they will talk to you about the increasing number of transistors on a chip and all the science underlying the ability to keep increasing, doubling the number of transistors every 18 months or so.

But if you ask an economist to describe to you the rise of semi-conductors, they won't talk about transistors on a chip. Instead, they'll talk about a drop in the cost of arithmetic. They'll say what's so powerful about semi-conductors is they substantially reduced the cost of arithmetic.

It's the same with AI. Everybody is fascinated with all the magical things AI can do and economists, what they bring to the conversation is that they are able to look at fascinating technology like artificial intelligence and strip all the fun and wizardry out of it and reduce AI down to a single question which is, "What does this technology reduce the cost of?"

And in the case of AI, the recent economists think it's such a foundational technology and why it's so important, it stands in a different category from virtually every other domain that we see today, is because the thing for which it drops the cost is such a foundational input, we use it for so many things, in the case of AI, and that's prediction.

And so why that's useful is that as soon as we think of AI as a drop in the cost of prediction, first of all it takes away all the confusion of, well, what is this current renaissance in AI actually doing? Is it Westworld? Is it C3PO? Is it hell? What is it? And really what it is, it's simply a drop in the cost of prediction and we define prediction as taking information you have to generate information you don't have.

So, it's not just your traditional form of forecasting, like taking last month's sales and predicting next month's sales, it's also taking for example if we have a medical image, say we're looking at a tumor. So, the data we have is the image and what we don't have is the classification of the tumor as benign or malignant, and the AI makes that classification. That's a form of prediction.

And so when something becomes cheap, from Economics 101, most people remember there's a downward-sloping demand curve, and so when something becomes cheaper that means we use more of it. And so in the case of prediction, as it becomes cheaper we'll use more and more of it. So, that'll take two forms. One is that we'll use more of it for things we traditionally use prediction for, like demand forecasting and supply chain management. But where I think it's really interesting is that when it becomes cheap that we'll start using it for things that weren't traditionally prediction problems. But we'll start converting problems into prediction problems to take advantage of the new cheap prediction.

So, one example's driving. Driving, we've had autonomous cars for a long time, or autonomous vehicles, but we've always used them inside a controlled environment like a factory or a warehouse. And we did that because we had to control the number of, think of it as the, "If, then," statement. So, we have a robot, the engineer would program the robot to move around the factory or the warehouse and then they'd give it a bit of intelligence. They'd put a camera on the front of the robot and they'd give it some logic, saying, "Okay, if somebody walks in front, then stop. If the shelf is empty, then move to the next shelf." If, then, if, then.

But you can never put that vehicle on a city street because there are an infinite number of ifs. There are so many things that can happen in an uncontrolled environment so you could never ... That's why as recently as six years ago experts in the field were saying, "We'll never have a driverless car on a city street in our lifetime." Until it was converted into a prediction problem. And people who are familiar with this new cheap form of prediction said, "Why don't we solve this problem in a different way and instead we'll treat it as a single prediction problem." And the prediction is what would a good human driver do?

So, effectively the way you can think about is that we put humans in a car and we told them to drive, and humans have data coming in through the sensors on our, the cameras on our face and the microphones on the side of our heads. And our data came in, we processed the data with our monkey brains and then we take actions. And our actions are very limited, we can turn left, we can turn right, we can brake, we can accelerate.

And the way you can think about it is think about an AI sitting in the car along with the driver and what AI is trying to do is, it's doesn't have its own input sensors, eyes and ears, so we have to give it some. We put radar, camera, light, around the car and then the AI has this incoming data. Every second it's got data coming in it tries to predict in the next second what will the human driver do?

In the beginning it's a terrible predictor, it makes lots of mistakes and from a statistical point of view we can say has big confidence intervals, it's not very confident. But it learns as it goes and every time it makes a mistake, it thinks the driver's about to turn left but the driver doesn't turn left, it updates its model. It thinks that the driver was gonna brake, the driver doesn't brake, it updates its model. And as it goes, the predictions get better and better and better and the confidence intervals get smaller and smaller and smaller.

So, we turned driving into a prediction problem. We've turned translation into a prediction problem. So, that used to be a rules-based problem where we had linguists with many rules and many exceptions and that's how we did translation. Now, we've turned it into a prediction problem.

I think probably the most common surprise that people have is, we have a lot of HR people that come into our lab and they say, "Hey, we're here to learn about AI because we need to know what kinds of people to hire for our company for a manufacturing or a sales or this or that division. Of course it won't affect my division because I'm in HR and we're a very people part of the business and so AI's not going to affect us."

But of course people are breaking HR down to a series of prediction problems. So, for example the first thing HR people do is recruit, and to recruit they essentially take in a set of input data like resumes and interview transcripts an then they try and predict from a set of applicants who will be the best for this job.

And once they hire people, then the next part is promotion. Promotion has also been converted into a prediction problem. You have a set of people working in the company and you have to predict who will be the best at the next level up job.

Then the next role they do is retention. They have 10,000 people working in the company and they have to predict which of those people are most likely to leave, particularly their stars, and also predict, "What can we do that would most likely increase the chance of them staying."

What I would say a black art right now in AI is converting existing problems into prediction problems so that AIs can handle them.”