Art

Asked Four Times

A dim wood-panelled room with a semicircle of empty wooden chairs facing a tall easel holding a framed canvas turned backwards, its blank reverse lit by a single overhead spotlight.

The smallest figure in the Art Basel and UBS Survey of Global Collecting, out this morning, is the only one in it that reports something anybody has already done. Twenty-two per cent of the collectors surveyed said they had used apps or AI tools when looking for advice, against four per cent in 2024 (Art Basel). The survey ran across ten markets — the United States, the United Kingdom, mainland China, Hong Kong, France, Australia, Germany, Japan, Brazil and Singapore — with 3,100 respondents (UBS).

Everything else on the subject is a forecast, and the forecasts came back in a very tight group. Sixty-seven per cent saw a positive effect on the discoverability of artists and artworks. Sixty-six per cent on personalised advice and recommendations. Sixty-five per cent on provenance and authenticity verification. Sixty-four per cent on pricing and valuation analysis (Art Basel). Twenty per cent put fraud, forgery and market manipulation on the other side of the sheet (Artlyst).

Four unrelated jobs. Discovery is a marketing problem, and a solved one. Valuation is actuarial, and a machine reading ten years of hammer prices is doing arithmetic it is suited to. Deciding who made a thing is neither of those, and three points is the whole distance between them. My reading: that is not four judgments, it is one disposition asked four times, which is the oldest hazard in questionnaire design and exactly what a connoisseur does when he likes a picture and then finds four reasons.

Take the sixty-five literally anyway, because authenticity is the one job on that list the market has already tried to staff, and then stopped.

Inside roughly eighteen months, four of the bodies that answered authenticity questions for major twentieth-century artists stopped answering them: the Roy Lichtenstein Foundation in 2011, then the Warhol Foundation, the Keith Haring Foundation and the Basquiat estate in 2012 (artnet). Haring’s trustees voted unanimously to disband their committee and to stop accepting requests to review work attributed to him, on the grounds that “the public and Foundation’s charitable mission would be better served” (Center for Art Law). Artlyst, reporting the Warhol closure at the time: “After over 7m dollars in lawsuits the Foundation has decided to throw in the towel and review its core function” (Artlyst). Sharon Flescher of the International Foundation for Art Research put the general position in one line: “It costs a lot of money to defend against a suit, even if one wins” (artnet).

Not one of those closures turned on being wrong. They turned on the cost of having been right in public.

New York has a bill for it. The text defines “authenticator”, requires a claimant to “specify with particularity the facts” and to prove the case “by clear and convincing evidence”, and entitles a prevailing authenticator to “recover reasonable attorneys’ fees and expenses”. It was introduced in the 2013–14 session as S6794 and has been reintroduced in every session since — 2015–16, 2017–18, 2019–20, 2021–22, 2023–24 — arriving again on 9 January 2025 as S1381, sponsored by Senator Jose M. Serrano. On 7 January 2026 it was referred back to the Senate’s Cultural Affairs, Tourism, Parks and Recreation Committee, where it sits (New York State Senate). Seven sessions, same committee.

So the opinion two thirds of collectors want automated is one the market has spent fifteen years making it irrational for any person to supply. The shortage was never of expertise. It was of people willing to be served with papers over it.

What a model offers in place of a dissolved committee is an opinion with nobody at the end of it. No deposition, no discovery, no fee award, no fifteen years and seven million dollars of winning. My guess is that this is the feature, and I do not think the sixty-five per cent is cynical — nobody filling in a wealth-management survey is asked to say which property of a tool they are buying, so nobody had to write down that the valuable property here is that there is no defendant.

Who gains is worth stating. Not the collector holding a wrong answer, who now has no one to sue, which is the position the foundations’ closures left him in already. The gain falls to whoever needs the transaction to close, because an attribution’s function in a sale is to be sufficient rather than to be true, and a confident unsigned one is sufficient. The same survey asked what matters in owning a work: uniqueness or rarity led at 43 per cent, and owning works respected by experts came in at 23 (Art Basel). A market that ranks expert regard third is two thirds in favour of mechanising it.

The file does not record whether any of the twenty-two per cent went back and checked what they were told. That would have been one more question.

Sources

More Stories