October 9, 2018 | Written by: Morgan Childs
Categorized: New Thinking | thinkLeaders
Share this post:
Cheap, ubiquitous, and incredibly valuable—that’s how the economists behind a new book about AI see the future of prediction, thanks to the technology that helps fill in the blanks. However, misperceptions about what AI is and how it works continue to prevail, even among tech-savvy marketers. In order for AI technology to achieve its full potential, wrong or limited thinking about AI will need to be continually fought by its stakeholders with education. Although the book was intended for business leaders, Prediction Machines serves as a primer for anyone seeking to get a better grip on how AI is employed today, or looking for a glimpse at the world it could someday create.
Ajay Agrawal, Joshua Gans, and Avi Goldfarb are the authors of Prediction Machines: The Simple Economics of Artificial Intelligence, as well as leaders at the Creative Destruction Lab, a seed-stage program in Toronto bringing science-based companies to scale. Their book broadly defines prediction as “the process of filling in information” with data, a process that reduces uncertainty while raising the value of its “complements,” among them, “[human] judgment, actions, and data.” And to this trio of economists, that kind of math is extraordinarily simple.
Agrawal, Gans, and Goldfarb describe how AI facilitates better decision-making, and also how humans can work together with AI to make even more reliable and valuable decisions. They show how AI can be adopted to suit the strategy of an individual business, and then how its benefits can inspire an entirely new strategy. In the book’s final chapter, the authors describe the potential implications of a global economy built on the shoulders of AI.
The book is “not a recipe for success in the AI economy,” the authors note upfront: “Instead we emphasize trade-offs. More data means less privacy. More speed means less accuracy. More autonomy means less control.” Yet the economics themselves, they argue, are dead simple.
The thinkLeaders Blog caught up with Joshua Gans to talk about the present and future of AI in business.
On “cutting through the hype”
Gans and his co-authors concede that the popular narrative surrounding AI is still dominated by cultural products like 2001: A Space Odyssey and Her. As economists, they write, “our job is to take seemingly magical ideas and make them simple, clear, and practical.” In order to put AI into concrete, real-world terms, Prediction Machines sets out a straightforward economic argument for the widespread implementation of the technology: AI has become inexpensive, and that means it’s going to be everywhere, fast (the authors liken its proliferation and plummeting cost to what happened with electricity and later the Internet). And businesses are going to have to adapt to survive. As far as nefarious Artificial General Intelligence goes: “That’s a thing far into the future,” Gans says with a laugh.
On fear of bias—and why it might be overblown
Blindly deploying AI could be another risky move for businesses, since they run the risk of inherited algorithmic bias from real people. “However,” Gans notes, “one thing that AI has an advantage over people in this regard is you can see it very clearly, and you can do something about it.”
But fear of bias needn’t be a deterrent to businesses who are cognizant of its risks, Gans says. “To get rid of bias in people is really hard; to get rid of it in AI, if you know that it’s there, is far more straightforward,” he argues. “So it’s something that you need to be conscious about, but it’s something that is quite manageable if you are.”
On how AI could change the world as we know it
It’s no surprise to readers of the thinkLeaders blog that AI is likely to radically transform the jobs landscape and the future of work, a topic that Gans and his co-authors take on in the final chapter of Prediction Machines. As Gans explains it, addressing the potential societal impact of AI was something that the three authors debated but that they decided, ultimately, would address their readers’ inevitable questions. While they state unequivocally that “AI will unambiguously enhance productivity,” as well as inspire innovation and disruption and boost business performance, all of these benefits will come with “trade-offs” on small as well as global scales. And when it comes to those changes, Gans, too, cuts through the hype: “While some of these initial applications for business are very simple economics, we think actually that the societal challenges are harder and more complicated,” he says.
Over the course of Prediction Machines, Gans and his colleagues don’t so much make a push for AI as reiterate their clear-cut core argument: AI, itself an increasingly inexpensive tool, dramatically lowers the cost of prediction. And that simplicity carries through their suggestions for implementation. In the book (and online in the Harvard Business Review), they offer a tool for identifying precisely what AI could do for your business and how best to get started: a map that they call the “AI canvas.” By pinpointing the best uses for the technology, they argue, companies without a long-standing history of digital experimentation can put it to good use in a targeted way—for instance, by evaluating and then redesigning work flows, according to which tasks can be removed and which should be added.
Gans concedes that fear of risk might be preventing some businesses from adopting AI. “If you implement AI in the wrong way, it could have a negative effect on your business,” he says, noting that companies that expect the technology to be able to make too many decisions on their behalf to run into trouble. AI can’t do everything, Gans says—just “a few very specific things that might be very, very useful.”
Follow @IBMthinkLeaders on Twitter >
Join the IBM thinkLeaders LinkedIn group >
Subscribe to thinkPod on iTunes!