The Knoedler Ghost and the Silicon Cure

Why the art market's future depends on infrastructure, not just intelligence

By Theo Johns

There's a story I keep coming back to — not because it's the most dramatic scandal in art market history, but because it exposes something most of us would rather not admit.

Between the mid-1990s and 2009, Knoedler Gallery in New York — one of the oldest, most respected galleries in the world — sold more than thirty paintings attributed to Rothko, Pollock, and Motherwell. Collectors paid millions. Experts examined them. Museums looked closely. And for over a decade, the works circulated through the market without serious challenge.

They were all forgeries.

What I find most instructive about Knoedler isn't the fraud itself. Fraud exists in every market. What I find instructive is how it worked — and how little would have to change for it to work again today.

The forgeries didn't succeed because the people involved were naive or incompetent. They succeeded because the information surrounding those artworks — their histories, their documentation, their verification records — was fragmented. Deliberately disaggregated. Each piece of the story lived somewhere different, and no one was responsible for holding it all together.

When information isn't organized and consolidated, even the best experts are forced to work in the dark.

The maze and those who know its routes

I've spent considerable time thinking about why the art market tolerates a level of structural opacity that would be considered reckless in almost any other industry. And I think the honest answer is uncomfortable: for insiders, the current system works.

If you know the market, you know the routes of the maze. You know which conservator has the relevant files, which shipper can speak to movement history, which lawyer's opinion carries real weight in a given jurisdiction. Knowledge travels through relationships. Trust is extended through reputation. Due diligence happens over introductions and handshakes, and the whole architecture depends on your proximity to the right people.

For those who already belong, this feels like connoisseurship. For those who don't, it feels like a closed door.

This isn't just an access problem, though it is that too. It's a structural vulnerability. Markets that rely on informal networks to route critical information aren't resilient — they're fragile. They can be gamed by anyone patient enough to learn the social codes and exploit the gaps between siloed sources.

Consider what it would mean if real estate operated the same way. The deed held by a private individual, no central registry, the title history the "word" of the current owner, and the bank, insurer, and buyer each working from a different version of events. In parts of the world where property markets work like this, they're treated as high-risk by global capital. Transactions don't scale. Confidence doesn't compound.

Real estate in Europe and North America works because it built infrastructure: title registries, standardized records, interoperable systems connecting banks, insurers, surveyors, and regulators around a shared source of truth.

The art market never built that backbone. And we are now perplexed by a market that has ceased to grow.

Technology is not infrastructure

Over the past two decades, the art world has cycled through a succession of technologies that were supposed to transform everything. X-rays and multispectral imaging. Digital marketplaces. Barcodes and RFID. Blockchain and NFTs. Now artificial intelligence.

Each wave carries genuine capability. And each wave, deployed into a fragmented data environment, produces more information without producing more certainty.

AI can analyze a painting with remarkable sophistication — but it still needs structured, reliable, primary data to provide meaningful findings. Scientific analysis can reveal extraordinary insights into a work's material history — but it can draw few conclusions without a verified body of comparative knowledge. Legal opinions can limit liability — but they rarely establish definitive truths and can sometimes obscure more than they clarify.

The problem isn't the tools. The problem is the foundation they're being asked to stand on.

I want to be direct about AI in particular, because the conversation has drifted toward hype in ways I find genuinely alarming. There is a category of AI — the generic large language model — that essentially guesses with authority. It has no reliable relationship with truth. It has a trained preference for plausibility and user agreement over accuracy. Researchers have documented how as few as 250 corrupted documents in a dataset of billions can plant a persistent backdoor for misinformation. How AI can be induced to repeat a false premise with full confidence simply because a user stated it with conviction. How "recursive pollution" — AI-generated content flowing back into training data — can cause errors to become increasingly ingrained over time.

If we deploy that kind of tool into an art market that already has a fragmented, partially falsified record, we don't solve the Knoedler problem. We industrialize it.

Used correctly, however, AI is among the most powerful instruments we have for consolidating knowledge at scale. The key phrase is used correctly — meaning: trained on verified sources, deployed to score the quality and origin of data rather than simply recite it, instructed to flag inconsistencies and audit the veracity of archives, designed to act as an ombudsman that corrects the historical record rather than reinforce existing errors.

This is AI as a power tool that amplifies expertise. Not a replacement for the scholar, the conservator, the scientist, the connoisseur — but a means of making their expertise scalable, searchable, and connected.

The difference matters enormously. And it requires something most technology vendors don't want to talk about: the data infrastructure has to come first.

A clinical chart for every artwork

When I describe what we're building at Authentify, I often use the analogy of a clinical chart.

When you enter a hospital, every interaction — every test, every diagnosis, every treatment, every specialist opinion — is recorded and connected to a single patient file. No one doctor holds all the information in their head. No one institution owns the record exclusively. The file travels with the patient, accumulates knowledge over time, and allows any qualified professional to enter the conversation with full context.

There is nothing equivalent to this in the art market today.

The identity of an artwork might live in a registry. The condition reports in a conservator's private archive. The provenance in a family's files. The scientific analysis behind an NDA. The insurance valuation in a broker's system. The movement history with a shipper who doesn't know they're tracking something important.

What Authentify is building is the connective tissue that brings these elements into a single structured environment — a persistent, verifiable identity for each artwork, one that follows international standards and can accommodate contributions from every relevant discipline without losing context or chain of custody.

Conservation reports. Provenance records. Legal documentation. Condition histories. Movement tracking. Environmental monitoring. Scientific findings. All of it coexisting within a framework that scores the quality and origin of each document, flags inconsistencies, and protects against selective curation — the quiet editorial practice by which inconvenient information disappears from a work's history.

The goal is not transparency for its own sake. There are legitimate reasons for privacy in this market, and a functional infrastructure has to respect that. The goal is defensibility — the ability to demonstrate, at any point, that a claim about an artwork rests on verified evidence rather than reputation alone.

When defensibility improves, value follows. And when value can be reliably documented, the market grows.

What the next decade decides

The art market is global in scope, but it still operates on infrastructure built for a smaller, more private, more relationship-driven era. Artworks move across borders constantly. Insurance values reach into the billions. New collectors are entering the market from regions that have no inherited trust in the existing social networks. Regulation is accelerating. The stakes are higher than they have ever been.

The question isn't whether the market will change. It will. The question is whether the change produces a more trustworthy, more scalable market — or simply a faster version of the one we have, with more sophisticated tools amplifying the same underlying vulnerabilities.

I believe the future will be decided not by who adopts AI the fastest, but by who builds the infrastructure that allows these technologies, and the experts behind them, to work together.

At Authentify, that is the work we are committed to: a unified, interoperable backbone; an identity layer that holds across jurisdictions and institutions; a documentation framework that eliminates selective curation; and an environment where scholars, insurers, regulators, marketplaces, and collectors can finally operate from a shared foundation of verified knowledge.

The Knoedler scandal happened because information was fragmented and the gaps were easy to exploit. The solution was never more expertise — the expertise was already there. The solution is an infrastructure where expertise has something reliable to stand on.

That infrastructure is what makes trust possible at scale. And trust, as anyone in this market knows, is the only currency that actually holds its value.

Theo Johns is a fourth-generation art dealer — the son of Bury Street Old Master dealer Derek Johns — and a specialist in European Old Master paintings with a particular focus on Italian works. He has held positions at both Sotheby's and Christie's in New York and London, and is an expert in provenance research. He joined Authentify Art as an early member of the team, bringing decades of market experience to a company building unified data infrastructure for the global art market. This article is adapted from his address at the TEFAF Maastricht Art Business Conference.