Another Good Intro To The Issues

Artificial Intelligence for Business, Kamales Lardi, is another good introductory book to the subject that suffers the same weaknesses as most of the other books I’ve read on the subject. Let’s start with the good. The book is a very accessible, easy to read review of what business management wants to know about artificial intelligence (AI).

At a high level, the review of AI is good, but ignore the details. I’m not sure if it’s because of confusion or a consultant’s need to use the buzzwords, but I’m not thrilled with some of the details. For instance, machine learning isn’t really limited to AI and was around in the business intelligence (BI) era. One thing good about the book is that it does mention BI, but doesn’t focus on the area I’ve mentioned in previous reviews. A lot being pitched for AI’s value has been done by BI for decades, including categorization and clustering. The difference is the volume of data that can lead to both higher precision and higher costs. It’s up to management to look at the necessary precision for a problem and decide if the ROI for BI or AI is better in each situation. Both AI academics and consultants want to push it, but remember it’s a key part of a modern solution and not a panacea.

A section of the book I really liked was chapter four, ethics. Ms. Lardi does a very good job covering both the concepts and examples. It’s the “must read” of the book. The chapter before that is ok, where it covers AI working with other modern technologies. Again, at a high level, it’s good; but the details are questionable. For instance, distributed ledgers are a key component of blockchain but that isn’t clearly defined. While distributed ledgers in supply chains and elsewhere are valuable, the examples I’ve seen have only uses that because consensus slows down real business processes and isn’t really needed. Again, I’m suggesting the reason that isn’t made clear is the “need” to push the blockchain buzzword.

Another mixed blessing is the chapter on the future of work. While the author does make an excellent case for massively increased unemployment, that case is mitigated with the usual apologia that “AI-driven automation does not substitute human labour completely … the human workforce will be able to focus on complex tasks.” As previously articles in this column, and those of plenty of other writers, have pointed out, there is a major problems with that thesis.

People can do the complex tasks in a process because they began as rookies with simpler tasks and moved up to the complex ones as they gained skill. If AI does the simple tasks, how are humans to learn the complex ones. Are business owner more likely to take new hires and spend significant time training them for the complex task or demand AI that moves upstream and allows them to replace all employees? Yes, that’s a rhetorical question.

The rest of the book is a good explanation of what’s needed to begin the process of expanding AI’s use in business and, of course, setting up the reasons why a consultant can help the reader. The first part is good. The second is neither good nor bad, just what is to be expected.

This is another book where the reader should always remember the author’s background and purpose. It’s far better than many coming out of academia, think tanks, and the blend of the two – people who came from academia, made a bunch of money at a startup without really understanding business, and who now think they know everything. The author is a consultant in the industry. The purpose of the book is to take her real life business experience, explain it to her market and, of course, drum up business. Remember that and it will be a positive read.