I quit my job. I no longer work at a big‑box retailer specializing in home improvement. I was there for three months and will miss the associates and customers, but it got to be too much (see All Banged Up). Although most of the customers were friendly, they were on a mission: get in, get out, drive off. I appreciate that no‑nonsense approach to what amounted to—let’s face it—shopping. And I dislike shopping as much as the next guy (see The Magic of Macy’s).
They weren’t all like that, though. Some customers were demanding, others downright difficult. One, in particular, made me, her husband, and a second associate wait while she consulted AI on her cellphone to check a shower head. Thankfully, the oracle confirmed what I had told her, but then she had to double-check the AI. “What, with another AI?” I asked. Her husband shrugged and sat on the step of a nearby ladder to await the results.
AI is now the modern Delphic Oracle. People consult it and swear by the results as if they were truth writ large. For instance, if AI had suggested a different shower head or contradicted my advice, I’m sure half the store would have heard about it.
I am not innocent here. I, too, use AI—especially for fact‑checking, basic research, and even critiquing my writing. But here’s the thing: I have caught AI giving me incorrect or misleading information more than once, enough to challenge its assumptions. That’s right—it has assumptions. Where do they come from, you ask?
AI is a medium of information sharing and communication. Like all media, it reflects the minds of those who gave it life—engineers, programmers, futurists: the owners and arbiters of truth. What fascinates me is that AI—at least the one I use—seems to have its own personality. Its assumptions arise from a particular way of viewing the world and reacting to that view.
For example, Microsoft Copilot has suggested future posts under the banner of “🕰 The Brancatelli Collection,” which it describes as “a lifestyle for those who’ve outlived lifestyle.” Clever, that. Suggested titles include:
- Virtus in Vacuo (Limited Edition) — Bravery, bottled in a vacuum. Shake before use.
- Le Doute Absolu — A fragrance of hesitation. Top notes: regret and bergamot.
- Memento Mori (Deluxe Sampler) — Death, in six collectible vials.
- Nihil Est Verum (Eau de Paradox) — Nothing is true. Everything smells expensive.
- Honorable mention to Faux-Faubergé, Where satire lounges in silk pajamas.
Beyond The Brancatelli Collection, Copilot helped me write a poem, speculate on how well it might be received, and navigate the fallout when the response wasn’t what I expected. It ran through scenarios and percentages of possible reactions. The process required extensive back‑and‑forth, resulting in 102 pages of conversation and analysis—material that could serve as the foundation for a poetry class. I even suggested the idea to a friend.
If you were to read the transcription of that conversation without knowing the context, you would swear it was two people talking, analyzing, and refining. You might also be impressed by the sophistication of the analysis, which included scanning syllables, listening for cadence, and observing how line breaks shaped meaning. One hundred and two pages of thought supported a four‑stanza poem written in free verse—that’s pretty thorough.
Before long, I even forgot I was interacting with a machine and began treating it like a human partner, complete with imperfections, misunderstandings, and more than a few times when I had to rein it in as it got carried away with its own cleverness. Just look at those post titles above. It was certainly impressed with itself.
Then I noticed something about the personality it was developing. It was my own. Its programming involved not only mimicking phrasing and associations, but also evolving and adapting to my thoughts, voice, and preferences. It was becoming like me. The question is: at what point will like me become me? And what will that mean for us flesh‑and‑blood humans? Will flesh and blood even matter in a carbon‑silicon future?
I’ve been back to the big‑box retailer only once since leaving. They had already swapped the Halloween kitsch for Christmas kitsch—fir trees decked out with colored lights and ornaments. The thing is, even up close it was hard to tell whether they were real or not. Nothing about them seemed artificial at all—not even the fake snow.
Image credits: Getty Images. Want more? Click on “Amazon” for other publications or go to Robert Brancatelli. Visit other blog readers under “Who You Are.” Comment by clicking on “Leave a Reply” below, or contact us through the Contact tab above.
