The Uncomfortable Truth About AI and Human Necessity

Like many people my age, I am looking after an aging parent. I consider it a privilege, but I am well aware that the years have smoothed over the more jagged edges of my mother's personality. She also is not dealing with a weird child but an eccentric adult who understands her as a woman and mother. One of her favorite sayings when I was growing up was, “Careful what you wish for.” The wishes of a child are often impulsive and do not account for unintended consequences, so it was wise guidance. Taking a look at some of the news from the technology world, it is clear that a few CEOs should have heeded that advice.

Economics News posted an informative video on how CEOs are beginning to regret their automation efforts with Artificial Intelligence. It comes shortly after a report by MIT that claimed 95% of Artificial Intelligence projects fail to increase revenue. The mass layoffs in technology and other white-collar work appear to be backfiring. I have a hunch what is happening.

First, Artificial Intelligence is expensive. It requires plenty of time to train, outrageous amounts of energy to power, and mountains of money to maintain a Large Language Model. It is why the major players like Microsoft, IBM, and Google are involved in its development. The smaller startup companies in the field are just leveraging the infrastructure of these larger firms and hope to be successful enough to be acquired by them. Companies outside the technology world fail to understand this reality until the bills arrive, and by then it is too late.

Next, business leaders are afraid of missing out on the benefits of Artificial Intelligence without understanding the way it might actually benefit their business. At first, the obvious answer was to help streamline call center and support functions in human resources. It soon became evident that these functions required judgment and human connection, which Artificial Intelligence does not do well. I was interviewed for a job by an Artificial Intelligence system, and the experience was so dehumanizing that I withdrew my application for the job at that company. It also explains why workers are standing guard at automated checkout kiosks at stores because customers need help using the systems and prefer human connection over automated voice prompts.

Also, many people deciding to implement Artificial Intelligence do not understand how their organization generates revenue or how work flows through the organization. It is why the Artificial Intelligence projects rarely make a return on investment because of this misunderstanding.

Finally, automation requires maintenance, and this is what Tesla discovered. Entropy is a universal process, and it impacts everything, including Artificial Intelligence systems. So people must spend time repairing, updating, and improving the automated systems which operate the business. If you skimp on maintenance, these systems break, and nothing is more expensive than a broken assembly line. It is why Tesla had to scramble to meet production goals. Elon Musk bragged that humans are overrated, but the uncomfortable truth is humans are necessary for automation to work correctly. Otherwise, you are building expensive and fragile systems that will require humans to save the day when customers demand results.

Business people are not dumb, so when they begin losing customers, revenue, or profits because Artificial Intelligence systems are not meeting expectations, the adoption of this technology will slow down.

I have been involved in technology and Artificial Intelligence since its emergence. The involvement inspires healthy skepticism in me, and I hope that my colleagues in the business community match my perspective. Otherwise, they will get what they wish for, and that might be a bad thing.