Stick with me and I promise I’ll loop this back to luxury and tie it up with a nice bow.

A classic science fiction trope is the idea of testing things that appear to be human to determine if they truly are. Star Trek: Deep Space Nine and the more recent Star Trek: Picard had their Changelings, shapeshifters imitating humans who could only be detected through an invasive or otherwise unpleasant test. John Carpenter’s The Thing had its alien that assimilated people, detectable by testing whether a blood sample would react to pain. Battlestar Galactica had its Cylons and Blade Runner its replicants, both types of androids that could, at their best, fool nearly anyone—raising questions of morality around what it means to be human and what we do or don’t owe something that’s apparently as intelligent and sensitive as we are.

We’re not even close to the level of Cylons and replicants yet, but we’re in the age of AI, and that means we’re in the age of AI slop, which Wikipedia defines as “digital content made with generative artificial intelligence that is perceived as lacking in effort, quality, or meaning, and produced in high volume as clickbait to gain advantage in the attention economy, or earn money”. It’s a founding principle of this newsletter to never use AI to write it, not because I’m a Luddite—far from it—but because I hope and believe that people still value stories. Stories about people, told by people.

A key challenge posed by AI slop is that generative AI is getting better literally by the month, week, and even by the day. So a tool that might have produced something that was obviously AI-generated six months or a year ago might produce something difficult to distinguish from anthropogenic (human-generated) today. That’s not a challenge for the people using AI to create content to distribute. From their standpoint, it’s just what they want. But it’s a challenge for the rest of us. When an internet source says, in effect, ‘You can trust me when I say that humans created this’, how will we be able to trust anything if it’s difficult or impossible for us to tell the anthropogenic from the AI-generated?

And if we’re going to talk about AI slop, we have to at least mention AI sludge. I think of it as what happens when there’s so much AI slop that it builds up in layers that make everything more difficult. We’ve been training AIs on anthropogenic content; what will happen if the Internet becomes covered in AI sludge, and new AI models train on the results of previous models? “Nothing good”, as the robot CASE says in Interstellar.

As consumers, AI-generated content is competing for our attention, and it’s winning that battle more and more all the time. We’re surrounded by content that is mostly or even completely untouched by human minds or hands. Short of observing the creation process, can any of us know that what we’re paying for actually came from a human? The best answer I can come up with at the moment is, “It’s difficult.”

Wait ‘til I get going! Where was I? Luxury. Yes, luxury!

However any of us may feel about it, luxury itself exists in the age of AI—and of AI slop and AI sludge. What’s to stop someone from—as an example—taking a set of silk scarf designs from Hermès, using that corpus to train an AI engine, using that engine to generate ‘new’ designs, sending those designs to a factory in a low-cost country, printing scarves on cheap silk entirely without human intervention, calling the results artisanal luxury, and marking them up by a factor of 10, 20, even 40? In fact, I wouldn’t be surprised if this is happening right now. One reason I wouldn’t be surprised is that I just did it myself—the first steps up through generating ‘new’ designs—and it took me five minutes:

Don’t like that colorway? This took another thirty seconds:

I showed these images to someone who is a fan of Hermès scarves. Their reaction was, “If I saw that design in a boutique, I wouldn’t be displeased. I’d be intrigued.”

Think of the hundreds and thousands of hours that must go into designing a scarf. The ideation and exploration. The sketching, drawing, and painting. The selection and testing of varying color palettes. The cycles of feedback and iteration.

Five minutes.

Does this scare you? It scares me.

But we could also see it as an opportunity. And crises often make for great opportunities. What is that opportunity?

(In Mandarin, 危机, or wēijī, means crisis. It used to be said that this word combined the words 危 (danger) and 机 (opportunity). More recent and authoritative scholarship tells us—as I recently learned to my disappointment—that this is at best an oversimplification, because 机 is better translated as critical juncture or turning point. But it’s such a useful way of thinking about things that I like to pretend that it’s true.)

A luxury firm could say something like, “To create our products, we use tools and machines, but we’ll never use a tool or machine that takes the place of the human mind.” That, to me, is a powerful statement. The problem is that anyone can say it. I say it about this newsletter. From my manifesto:

(Obviously that doesn’t include content which I specifically call out as AI-generated and include for commentary purposes, such as the images of the scarves above.)

In other words, while I use AI extensively in my personal and professional lives, including to help me conduct research for this newsletter, the words I write are mine and mine alone—always.

How can you, as a reader, believe that? It’s difficult with the written word like this. I think it comes down to establishing a relationship between you (both “you” individually, as in the person reading this, and “you” as in my entire audience) and me: a relationship in which my writings incrementally convey my personal and very human point of view, with all its quirks. That point of view can, should, and will evolve over time, or I’m not upholding my personal standards for flexibility and growth. But it should also have an underlying consistency, a through-line that I hope you see in everything I write here.

What if I weren’t a commentator on the luxury industry, though, but were instead making products in the luxury industry? Would that change how I could go about creating trust with my customers? Absolutely.

In this article, I’ve described the problem: AI makes it harder to believe the claims of luxury. Next week, I’ll describe one possible solution: the luxury boutique as a source of trust.

We haven’t yet reached the point where AI can convincingly simulate physical experiences, including not just sight and sound but smell, taste, and touch. That technology is decades away at the earliest. (Talk about science fiction tropes!) And there lies the opportunity:

In the world of luxury, physical boutiques matter more—not less—in the age of AI.

Physical boutiques offer forms of evidence that AI can’t reproduce. AI can generate product images, like those of the scarves above, but it can’t simulate the texture of the silk of a real scarf. AI can tell you how much a scarf would weigh, but that’s not the same as feeling how a real scarf drapes around you. AI can tell you what type of steel a mechanical watch is made from and how that steel is finished, but that’s not the same as feeling the solidity of the metal and how it sits on your wrist. AI can tell you what a ready-to-wear garment is made from and how it was constructed, but that’s not the same as feeling how it moves with your body.

This suggests a fundamental rethinking of the role of the luxury boutique. Not the boutique as Ozymandian theater, or as a prettier checkout counter with Champagne and sparkling water on offer, but the boutique as a place where provenance, material, craft, and service can be tested in person. In other words, the boutique as a place to experience humanity in a time when AI makes that increasingly difficult.

The question isn’t merely whether boutiques still matter in the age of AI. It’s how they should be designed when trust itself has become part of what luxury must sell.

For now, what I would say is this:

If something is too good to be true, it’s probably too cheap to be anthropogenic.

In other words, that bargain is a bargain because its design and production were optimized to make it a bargain. AI slop designs (or artists not paid fairly for their work.) Low-quality materials. Mass-scale manufacturing. No artisanship. There’s a reason that human-designed, handmade silk scarves cost hundreds of dollars, even when they’re from maisons whose names aren’t Hermès.

Caveat emptor luxuriae.

Let the buyer of luxury beware.

Would you value a luxury object less if you learned that its design had been substantially generated by AI, even if the finished product looked beautiful and was well made?

What do you think luxury maisons should (or must) disclose about the use of AI in the design and production of their products?