Power and Prediction (and the winners and losers when they shift)

This book review was not written by ChatGPT or any other bot. Apparently, that may be rare going forward . . .

Just in time for the holiday season, the three wise men from the University of Toronto’s Rotman School of Management ( Ajay Agrawal , Joshua Gans , and Avi Goldfarb ) brought us a treasure trove of insight and analysis about AI adoption and the transformative impact it will like have on companies, industries, hospitals, governments, and more.

“Power and Prediction: The Disruptive Economics of Artificial Intelligence” picks up where the trio’s 2018 bestseller? “Prediction Machines” left off. It’s absolutely worth a read for business leaders and others who are interested in how AI will likely impact our economies and societies in the coming years.

One reason to read Chapter 1: You’ll learn a useful framework for how AI may go about reconstructing companies and industries in very similar ways to how harnessing electricity led to reconstructed companies and industries. Nugget #1: “Our task here is to light your way to anticipate who may gain and who may lose power as AI systems develop and are adopted.”

One reason to read Chapter 2: You’ll orient yourself in what the authors refer to as “The Between Times.” AI’s here, but it’s still viewed by many as niche and has barely scratched the surface in terms of how impactful it may well be in the reasonably near future. Nugget #2: “AI prediction is already having a system change effect in the innovation process.”

One reason to read Chapter 3: You’ll learn what kinds of companies tend to be the most successful early adopters of AI technology and why. Nugget #3: “The enemy will go beyond the training data, and peacetime data will be of little use.”

One reason to read Chapter 4: You’ll gain an understanding of how and why AI offers much more value to organizations that have a lot of decisions to make (in real time) versus organizations that are mostly run according to standardized “rules” (i.e., pre-loaded decisions). Nugget #4: “The function of AI is to provide better prediction, which essentially means that you have the information you need to make better decisions.”

One reason to read Chapter 5: You’ll never think about airports in the same way again. Nugget #5: "If your business is to provide a way to help people when they wait for a plane, what’s the chance you are going to ensure that they don’t have to wait for planes?”

One reason to read Chapter 6: You’ll be exposed to some of the ways in which AI can turbocharge personalization in both advertising and education. Nugget #6: “The most profitable aspect of the AI would be to ratchet up advertising to customers who were on the fence between the free and the paid versions.”

One reason to read Chapter 7: You’ll see how a system that has too many rules may be too slow to implement predictive tools that would make the system more effective. Nugget #7: “A new system may be so disruptive that you may need to start using it in a new organization, where it can grow organically, rather than trying to adapt to it in existing organizations.”

One reason to read Chapter 8: While people tend to think about potential cost savings from AI, more sophisticated players (such as some of the biggest tech companies) focus more on the value that AI can unleash. Nugget #8: "Chatbots are playing a bigger role in customer service, and machine translation is getting an increased share of that activity.”

One reason to read Chapter 9: You’ll see how automation of hypothesis generation might lead to enormous advances in medical research and recommendation-engine development. Nugget #9: “If we had to point to an area where AI has the most potential for transforming the economy, it is well upstream from most ordinary business activities: in the system of innovation and invention.”

One reason to read Chapter 10: There’s a great discussion riffing on the theme that top-down innovation is much more dangerous to incumbents than is bottom-up disruption. Bottom-up disruption is much easier for incumbents to co-opt. Nugget #10: “The Blockbuster case is, of course, a dramatic example of both the failure to change in the face of a new technology and also how internal forces prevented that change before it was too late.”

One reason to read Chapter 11: AI doesn’t make decisions, but – as AI becomes more and more sophisticated – it sure *looks* like it’s making decisions. You’ll become more sophisticated in understanding where decisions are actually made in different systems. Nugget #11: “What Amazon was doing was measuring worker performance, using an AI to evaluate whether that performance was a cause for concern, and then having a human manager decide what to do.”

One reason to read Chapter 12: You’ll read about how early access to feedback data can give first movers on AI a possibly insurmountable advantage. Organizations that wait to adopt AI will face higher and higher barriers to entry as we go through time. Nugget #12: “Learning comes from data, and so first-mover advantage comes from a feedback loop in that data.”

One reason to read Chapter 13: I’ve got to admit that I wasn’t necessarily expecting to read about Michael Jordan and the Chicago Bulls. But it is a memorable example of how judgment is affected when prediction is decoupled from judgment. A variation on that theme begins with Vinod Khosla 's conjecture about potential interplay between AI, radiology, and litigation. Nugget #13: “In the end, [Michael] Jordan ‘took the pill” and got back in the game, albeit with significant time restrictions that [Jerry] Reinsdorf imposed.”

One reason to read Chapter 14: As multiple systems transition from deterministic to probabilistic, missing data means uncertainty, and we can expect to see enormous impacts in everything from self-driving cars to international law interpretation and enforcement. It’s more and more about “thinking in bets” as poker star Annie Duke put it. Nugget #14: “These changes may be disruptive, creating discord as they are implemented.”

One reason to read Chapter 15: You’ll read about a situation where University of Michigan’ professors Eric Schwartz and Jacob Abernethy built an AI that transferred important decision-making authority from the legislative system to the judicial system (and seems to have been a force of good in this case as far as I can tell). Nugget #15: “The losers may be whole parts of organizations – such as Blockbuster’s franchisees that were against streaming videos.”

One reason to read Chapter 16: You’ll get a nice look at how synchronization problems in a system (e.g., a supply chain) can be?alleviated through either coordination or modularity. Modularity has important limitations but can be useful in situations where coordination is not feasible. Nugget #16: “Amazon has a modular organization that has allowed it to slot better AI prediction into recommendations that minimize the impact on the rest of the organization.”

One reason to read Chapter 17: There’s a good walk-through of how automating process allows businesses to create more value for customers through customization. And there’s some good analysis of where the insurance industry is and may be heading with its AI adoption. Nugget #17: “You want a system that can automate both the prediction and also the delivery of the products to the customers.”

One reason to read Chapter 18: The authors remind us that we’re still in the “early days of the In Between Times,” but suggest that we can expect to see access to information and allocations of decision-making authority within organizations will present interesting challenges for business leaders in the near future. Nugget #18: “System innovation requires disruption, which carries with it changes in the distribution of power with winners and losers.”

Since I first read Alvin Toffler’s predictive masterpiece “Powershift” (in 1997 ?!?), I’ve periodically wondered when I would next come across another book of comparably relevant prescience.? Is “Power and Prediction” a book that’s going to give us a whole lot of visibility into how the next two or three decades are going to play out? It’s too early to tell for sure, obviously, but I think it may be. We’ll know a lot more in 3 or 4 years. There’s a lot in this book, and it’s well worth your time and attention.

#businessleaders #chatgpt #machinelearning #AIadoption