Almost every week, one technology vendor or another tells me that their AI product will free workers from mundane jobs and enable them to do more exciting work. And every time, I respond with, “That is simply not true” (though more diplomatically). As AI works its way through blue-collar jobs, lower-paid white-collar jobs, and now into higher-paid professions, that sales pitch falls flat. In theory, AI automation can free workers from the mundane and create new and more exciting jobs. Similarly, AI automation can help fix inefficiencies, reduce bottlenecks, boost accuracy, and improve jobs. But in reality, that seldom happens. Processes are not fixed; rather, workers’ roles are simply replicated, and a human worker is replaced by a digital worker.
Take blue-collar jobs. In just one example, AI-driven robots are starting to make a significant impact on the world of construction. Robots like the Sam100, which can lay bricks at three times the speed of a human, are replacing human bricklayers, many of whom are not thrilled. AI that automates sales and call center work has the same effect. Calls are increasingly being handled by digital agents, reducing the need for human agents.
Myth of Job Creation
Another argument I hear from tech leaders is that AI will actually create more jobs. This argument is delusional. If you automate a job, then by default, that job is removed from the pool of jobs available. Even if you move the workers impacted to other duties, the people you would have hired for that original work no longer need to be engaged. As technologists, we need to get real about the impact of AI and be honest about the effects on peoples’ jobs and futures. Nobody outside of the self-delusionary tech bubble believes one word of these pro-AI automation arguments anyway.
In more than 25 years of advising and consulting organizations around the world, I have never seen a business case that involves the altruistic premise of moving workers to better jobs. Business cases are uniformly focused on more efficiency, lower costs, and—most commonly—headcount reduction. However, it is worth noting that 7 out of 10 IT projects fail or fall short of expectations because the expected efficiencies, cost saving, and headcount reductions do not materialize. One could say it’s a case of business leadership overestimating its own value and understanding of the business while underestimating the complexity of its business realities and, in turn, grossly undervaluing its employees.
The work of the U.S. government’s Department of Government Efficiency (DOGE) team, whose openly stated goal is to automate government jobs and lay off workers, is a high-profile example of AI and automation being used to replace workers, not augment or improve work. Its approach is not novel for a business, but is unusual for government agencies. Whether DOGE will deliver the promised massive cost savings and efficiency gains is yet to be seen, as it’s early days. But such brute force approaches seldom do.
None of this is to say that Deep Analysis is against AI. In fact, Deep Analysis is a proponent of using AI for business automation. We created an AI training course; we have advised countless organizations; and we have even written a bestselling book on the topic. But we believe AI can and should be used not expressly to replace humans, but to completely rethink how we work.