I recently had a great conversation with Les Shute. He is a go-to leader in the world of artificial intelligence. My perception has been that he is part of a very small slice of the world that has both interest and appreciation for the implications of AI.
He corrected me, sharing his perspective that he is part of a much larger group than I perceive, but admitted that there is an exceedingly small intersection of that group and the workers' compensation community. In other words, he is part of the very small group of AI experts in our little corner of the world.
Mr. Schute is invited to speak at conferences regarding AI, and I have been fortunate in his patience and persistence in explaining various elements to me. The fact is that AI moves with alacrity and stealth. It is evolving in various ways, organic and calculated, patent and latent. We might debate the use of "evolving," more on that below.
We are in the midst of changes in the broadest context. AI is being integrated into every aspect of our world. See The Trough of Disillusionment (December 2024). It has already become ubiquitous, and yet its influence is growing.
Every time I speak with Horace Middlemier* about anything, he immediately asks, "Shall we see what AI thinks?" Every aspect of Horace's intellectual life has become infected with at least the specter of AI. But for some, more than a specter, it has become a singular focus that is as much a distraction as help. You can pick condiments for your burger without seeing what AI thinks.
That inclination toward ubiquity will be fed by the ease of AI access. Mr. Shute caught me flat-footed when he invited me to interact with chatbots that he uses in his day-to-day and in presentations. I touched on this in Steering Wheels or Paranoia (October 2025), and delivered some compliments. I was genuinely impressed with the bots and the conversational nature of the interaction.
I asked the bot some probing questions, though I was unprepared for the opportunity. I initially thought that it might be a prank, but there were soon signs that my dialogist was a machine. For one, there was a persistent tendency to be deferential and conciliatory. When I disagreed at one point, it apologized for having expressed a contrary conclusion.
I described that experience to Bob Wilson and he immediately invited me to join him on a video call and to speak with his chatbots. I was struck by the similarities but also with the differences. I perceive that the "personalities" of these tools have likely been influenced by the predilections, emotions, and personalities of their creators.
As the world rapidly adopts these tools, there will be evolution. In my conversations with the chatbots, I have learned that they are verbal (and optical) manifestations, and they are speaking the same output that you could harvest in a written response to a prompt fed to an large language model (LLM). The chatbots are dependent upon the language, context, and content on which they were trained.
Those chatbots will adapt as we proceed based on the language, syntax, slang, and emotion of conversations in which they engage. They will be learning from your call to customer service, focused on your stated desires, unspoken concerns, and even personal misinformation. They will learn, as humans have, to discern what is really needed or wanted despite the human caller's inability to precisely articulate it.
In short order, these chatbots will become increasingly difficult to distinguish from real persons. That will benefit the customer service element and lead to lost jobs. See Did They Warn You (November 2025).
I recently received some hate mail from a company with which I have done business for years. They did not intend hate; they had merely drafted an incompetently misworded letter. As I strove to deal with their misfeasance, I discovered that no one in authority there would engage. Those employees, I was assured, were "not customer facing."
Instead, I faced a parade of clerks with officious titles, each as unhelpful, inept, and conciliatory as the last. This company will undoubtedly evolve quickly to chatbots, as their intent is to frustrate communication, blunt criticism, and strive for placation over service. Poetically, they are a "service" company that has utterly lost sight of both the service and its delivery. They will persist with advertising, jingles, slogans, and superficiality. And they will do so because it is largely sufficient in the eyes of their customers.
But, this presents serious threats of misplaced trust in more personal arenas. See Steering Wheels or Paranoia (October 2025). As we become increasingly comfortable in their business deployment, humans will need to guard against complacency in their use in more personal settings.
For more on my AI and tech musings, there is a complete list on my website.
*Horace Middlemier is not a real person but a literary foil and tool. Any resemblance to any real person, living or dead, is unintended, coincidental, and happenstance.
