Surveillance pricing, an art market staple, is about to hound us everywhere
AI is turning a simple practice among dealers into an inescapable monstrosity of retail
A wall-mounted surveillance camera whose base forms a makeshift ‘O’ in painted text that reads ‘I SEE YOU.’ Photo by Aleksei Sabulevskii on Unsplash
One of the first lessons you learn from working in the gallery sector is that the price of an artwork often depends on who wants to buy it. The more influence (and often, wealth) a person has in this weird industry, the bigger the discount they tend to get from dealers. This means the price of a painting isn’t just a function of supply and demand; it’s also a function of a particular client’s identity, standing, and track record.
If this practice sounds infuriating and anti-democratic to you too, I can’t say you’re reacting irrationally! But I regret to inform you that an AI-powered version of it is about to become inescapable in nearly all phases of your life as a consumer. In fact, the shift to so-called surveillance pricing is already underway—and barring a grand reversal of fortune, it’s poised to make the gallery sector’s crude preferential pricing look quaint in comparison.
Counterintuitive as it sounds, there is a real business logic behind dealers’ charging lower prices to more powerful people than to less prestigious competitors.1 Remember, the art trade is first and foremost a social network. Most people in it want to buy works by the same artists that have already been collected by other people they respect, like, or envy, especially if the market value of those other people’s collections has gone intergalactic over time.2 This phenomenon means one of the highest-upside moves in a dealer’s repertoire is to selectively apply sizable discounts to pieces offered to industry tastemakers, because closing those few deals hugely strengthens the odds of selling much more of that same artist’s work both now and later in their career. It’s a short-term sacrifice for longer-term gain.
Of course, dealers can’t play favorites successfully without having a good sense of whose approval actually has the power to nudge the market. Gathering that knowledge requires softly surveilling not just the collector pool but also the larger patronage and influence ecosystem, then pricing accordingly when a noteworthy collector expresses interest. Because the process is kinda rudimentary, it provides a useful on-ramp to conceptualizing the algorithmic hellscape that commerce in general is en route to becoming, too.
For a preview, look no further than Delta Airlines, which recently became the first major carrier to announce it is using an AI model to offer different prices for the same seats to different customers depending on their individual data histories. Those last six words are crucial. Surveillance pricing “is different from old-fashioned dynamic pricing—where prices go up or down for everyone depending on market demand—and much creepier,” Charlotte Cowles writes for The Cut. Instead, the goal is to use advanced algorithms to charge fees that are “microtargeted to customers based on their unique profiles,” including their age, ethnicity, location, income, credit history, browsing history, and spending habits all the way down to, say, their proclivity to “order a specific thing at a certain time of night,” she adds.
Glen William Hauenstein, Delta’s president (and not, despite his name, either the antagonistic dean in a bawdy college sex comedy or the corrupt C-suite villain in a Robocop sequel), told investors during a July 10 earnings call that the company had already applied surveillance pricing to 3% of its domestic flights over the preceding six months. The scheme delivered, in his words, “amazingly favorable unit revenues” that convinced him and the other executives to try to up the technology’s usage to 20% of Delta’s tickets by the end of 2025.
It should go without saying that the excruciatingly dorky phrase “amazingly favorable unit revenues” means that the initiative made a killing for Delta. Which means, in turn, that the average air traveler was the victim. And the massacre is fast expanding well beyond a single predatory airline.
Surveillance seer
At least one person saw this mess coming years ago. Surveillance pricing is the key component of a larger framework laid out by Harvard sociologist Shoshana Zuboff in her 2019 book The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power. The core idea is that we’ve entered an era in which corporations relentlessly collect data about who we are and what we do, both online and off, to build fast-twitch predictive models that can extract the maximum possible retail value from each of us at all times.
One reason pricing in the dealer sector qualifies as proto-surveillance capitalism is that it relies on an imbalance between buyers’ and sellers’ intel on the market. Even though price and sales data are vastly more prevalent and accessible now than in earlier decades, dealers still know way more about their inventory, as well as the supply and demand for it, than even the best researched collectors.
There’s nothing wrong with that disparity in principle—until or unless, that is, some dealer presses their advantage far enough to outright exploit a client. (It’s relatively rare in my experience, but it happens.) Similarly, Zuboff writes that the danger of surveillance capitalism is that it “operates through unprecedented asymmetries in knowledge and the power that accrues to knowledge.” She continues:
“Surveillance capitalists know everything about us, whereas their operations are designed to be unknowable to us. They accumulate vast domains of new knowledge from us, but not for us. They predict our futures for the sake of others’ gain, not ours.”
Early evidence suggests this information asymmetry is already wreaking havoc on consumers wherever surveillance pricing rears its head. An FTC investigation in the final months of the Biden administration found that commercial intermediaries for around 250 retailers “can show higher priced products based on consumers’ search and purchase activity,” such as by frontloading more expensive baby thermometers in the online search results of anyone whose data profile pegs them as a new parent.3 On a smaller scale, market research by Consumer Watchdog, a California-based nonprofit, “found that the best deals were offered to the wealthiest customers—with the worst deals given to the poorest people, who are least likely to have other options,” Irina Ivanova at Fortune reports.
The path forks
Despite the few noteworthy similarities, however, it’s important to draw out the serious differences between the soft surveillance pricing in the gallery sector and the nightmare version Zuboff warned we’ve been sliding toward ever since Google started scraping data from users’ Gmail inboxes and selling it to advertisers.
For starters, there’s a meaningful ethical divide between, on one hand, a dealer knocking an extra 20% or 30% off a new painting by a promising artist because it’s going to an influential collector and, on the other hand, an airline price-gouging someone whose search history suggests they need to take an urgent flight to see a gravely ill family member. It isn’t just that the former scenario revolves around a discretionary purchase and the latter revolves around a human need. It’s also that, unless we’re dealing with an absolute crook of a gallerist (again, rare but hardly unheard of!), the least favorable price someone can get is still just the original list price for the artwork. In contrast, there’s nothing stopping an airline from jacking up a particular traveler’s ticket price to multiples of what it would be in a more neutral situation.
A chasm also separates the aggression and precision of AI-enabled surveillance pricing from the ad hoc discount regimen of the dealer sector. The former is digital, endlessly dynamic, and hyper-specific; the latter is analog, relatively static, and ultimately pretty simple. An advanced algorithm can fine-tune the prices of thousands of products by a little, a lot, or anything in between in response to numerous quantifiable factors liable to shift from moment to moment for a nearly infinite number of customers (as the FTC alleged Amazon did in 2023). A dealer, in contrast, can’t do much more than assign round-number percentage discounts to primary-market artworks in their own inventory based on their read of a small handful of clients’ relative amount of juice in the industry—a trait that usually only changes every few years, if at all.
I doubt these gaps will ever close much. True surveillance capitalism doesn’t just require mountains of personal data. It also requires markets with prices that are, and have always been, allowed to move freely. The primary art trade gets disqualified on both counts. Collectors don’t browse, let alone buy, enough art through trackable digital platforms to build robust quantified profiles in the first place. Primary prices are also controlled by dealers, who use an abstract (and in some cases, improvisatory) calculus to try to prevent their artists’ individual markets from escalating too far too fast—and at the same time, to ensure their prices never visibly fall, either.4 The combination of these two dynamics makes the market all but impervious to Zuboff’s worst fear.
That fear, by the way, runs deeper than customers being bled dry by asymmetrically informed price gouging. She warns that, armed with their terabytes of data about the consumer population, corporations will actually be able to covertly influence our behavior in ways that will worsen the financial harm beyond what they could do by flawlessly anticipating our tendencies alone.
The gallery sector certainly deserves some tough love for the modest (at best) technological progress it’s made this generation. Nevertheless, at least its plodding pace of innovation and its commitment to IRL experiences should make it a minor refuge from the AI-powered retail torture chamber we’re headed toward. It isn’t much, but like a predatory pricing algorithm, I’ll take whatever I can get.
Although money and influence often go hand in hand, it’s important to point out that art dealers don’t necessarily treat them as equal. Here, a buyer’s status within the art world usually outweighs their pure net worth when it comes to pricing—a feature of the primary market that makes some obscenely rich neophytes want to throw furniture through windows, once they discover they can’t always get an artwork they want just by paying more than someone else.
There are all kinds of names for this principle. The most formal is “social proof,” but I hate using it because it got co-opted by pickup artists in the 2000s. I’d advise you to choose between more familiar terms like the herd mentality or the bandwagon effect to prevent people from thinking you look up to a statement hat-clad goof who calls himself Mystery.
That the FTC’s investigation was promptly shut down shortly after a certain someone got inaugurated this past January will no doubt deliver you the kind of full-body shock usually only achievable by being pushed through the surface of a frozen lake.
There are workarounds to this rule in practice, including offering deeper than normal discounts or, since an artwork’s price typically relates partly to its size and materials, motivating artists to make works that are smaller or less prestigious, like drawings instead of paintings. But on a like-for-like basis, it’s extremely rare for one artist’s list prices to shift noticeably, no matter their career trajectory.
