For their first weekly edition, the revived London Evening Standard plans to let the notorious art critic Brian Sewell loose on the National Gallery’s new “Van Gogh: Poets and Lovers” exhibition. This marks a rare appearance from Sewell: he hasn’t had any by-lines in nine years, mostly on account of being dead.
The Standard’s owners have not discovered the art of reanimation. They’ve announced, instead, that they will use AI to produce a review in the style of Sewell’s existing corpus. The ironies of this stunt are carefully calculated: the critic was fascinated by forgeries and counterfeiters, and once admitted to actually painting over a faux Hogarth and selling it to the Tate. He made a point of declaring himself ‘unmoved’ by Young British Artists (distrust of the new, perhaps?). It is also funny to dispatch a robot to review Van Gogh, a man so overwhelmed by the force of human emotion that he cut off his own ear. Sewell’s estate are said to be ‘delighted’ with the arrangement, which the Standard claim will be a one-off.
Readers of the slimmer, rarer ‘London Standard’ (as is the rebrand) are being scammed. It’s not so much an ethics issue as an exclusivity one – you wouldn’t write a press release to say you’re going to print the previous night’s lottery numbers. Anyone with internet access can ask a chatbot for a Brian Sewell-style review. When I tried getting ChatGPT to write about Van Gogh as Sewell would, it produced a credible imitation of his weary rhythm. But a real-world contrarian could not have written the piece: it was completely predictable, littered with complaints about the artist’s inflated reputation and dodgy brushwork. I tried again with Jeff Koons (‘an art world charlatan, peddling polished detritus’) and Velazquez (‘subtle yet profound’).
Actual critics, unlike neural nets, do not simply extrapolate from one bit of culture to another. They have biases, rivalries, sexual fetishes, points of personal interest, and adverse experiences – these colour their responses to art and to life. Large language models can come up with serviceable copy, but you’d never get one to recreate the set pieces that make the British media such a notorious attraction. An AI theatre critic would never stray off-course to write about its night in London’s red light district, as did Lloyd Evans in the Spectator this April; you couldn’t get robo-columnists to drop the act and cuss each other out ad hominem, as in the infamous 1993 Modern Review Fax War between Julie Burchill and Camille Paglia. To err is to be human.
The demon columnists – feared, venerated, and deeply flawed – are on their way out. Art history courses discourage Sewell’s brand of sweeping conservatism, and for many young people interested in culture, acerbic reviews have given way to drawn-out video essays, which make more effective containers for advertising. The AI wave is ultimately empty – but the Standard’s stunt is valuable. It will show us what we are losing.
