“Working with AI”, read it for the case studies

This is going to be a soap box item. “Working with AI,” Thomas H. Davenport and Steven M. Miller, is a book from MIT Press that has some good points, but overall loses out. As the title says, this book has a good variety of case studies, but the highest point they try to make is not correct.

Let’s start with the good. As mentioned, the first two-thirds of this book is a series of fairly short case studies that show the breadth of artificial intelligence (AI) adoption in the business world. AI is an invaluable tool and is increasingly being used in many ways that affect businesses and citizens on a day-to-day basis. For those who want to understand the rapid spread of technology, this book will help you see just how adaptable AI is and how it is transforming knowledge and processes.

It’s that last part, however, where the authors fail, and they most likely do so on purpose. The repeated trope that AI won’t destroy jobs is used in a way that academics provide for management to use in their own corporate messaging. Messaging, not reality. The AI ​​revolution is not an industrial revolution, and economies are not where they were in the 18th century.

The best way to talk about what will happen to jobs is to use one of the case studies and then a polemic part later in the book. A case study is about how underwriting has changed. AI is being used for the simplest underwriting cases, arguing that it will help reduce boredom for experts and allow them to focus on complex cases. A knowledgeable employee, however, wonders how they will train the next generation of underwriters, since humans will no longer look at the easy, the policies that people learn to prepare for complex cases. It’s a valid point, but it’s short-term.

The authors repeat a discussion of that challenge in a higher-level policy chapter in the last third of the book. While they propose that schools teach foundational knowledge for all affected fields, a truly unrealistic concept, what they fail to do is acknowledge the reality of AI. The same owners and executives who are hollowing out public school systems are seeing what current AI systems are doing as just a start. So are programmers and system designers.

The goal of the people who build and buy systems like the ones mentioned in the case studies is to replace skilled workers as well. Then there is no need for the high cost of personnel that reduces the bottom line. You don’t need entry-level workers if you don’t need experts either.

I love AI, but I understand that it’s a huge change in the way our societies will work, and not enough attention has been paid to it. The sooner we look at how the definition of work is changing, and how to make real changes to our education system to help the dwindling middle class adapt to those changes, the better chance we have of integrating AI and defending and improving our society.

The academics who talk to executives are doing a good job of showing the benefits of AI and how adoption is accelerating. However, it seems that most of the academics writing the books I’ve reviewed either don’t really understand how AI is going to change both business and society or are intentionally ignoring the topic in order to sell books that can be put on their resumes.

This book is an example of that. He does a great job of describing the breadth of tactical AI deployment, and he does a poor job of looking at the strategic goals and impacts of AI.

Leave a Reply

Your email address will not be published. Required fields are marked *