Dior's hand-drawn holiday film has turned into an argument about AI, and it carries a lesson for every DTC brand heading into Black Friday. When feeds are flooded with machine-made content, "made by humans" becomes a claim that shoppers will test, and a brand that can't back it up pays for it.
What actually happened at Dior
The short film, directed by Sylvain Chomet, was widely praised as hand-drawn. Then viewers spotted what look like generative-AI artefacts. Jonathan Anderson told WWD the film is entirely hand-drawn, while a film-studio CEO claims AI was used and that the behind-the-scenes video was staged. As of Fast Company's report, Dior had not responded to the staging claim. The writer's argument, which we agree with, is that AI itself is not the real issue. Disclosure is.
We should be clear about what we know. This is one report, the claims are unresolved, and we are not judging who is right. The useful part is the shape of the problem: a craft story only works while people believe it, and the behind-the-scenes footage that was meant to prove it became the thing under suspicion.
Craft is becoming the signal, well beyond luxury
Dior isn't alone in leaning on human effort. Glossy reports that Levi's, Prada, Dior and Oribe have all made time-intensive hand-drawn or stop-motion holiday work, often hiring known animators and showing their process. Experts quoted there say visible effort reads as quality, and that shoppers have grown wary of AI imagery after criticism of recent Gucci and Valentino campaigns. Fast Company has also linked Prada's hand-drawn campaign to the same shift.
The same instinct is showing up lower down the price ladder. Beauty Independent surveyed 11 beauty and wellness founders and found them moving from polished posts to educational, behind-the-scenes and candid content, with formulaic "problem, solution, buy" stories reading as less credible. One brand says founder content now drives most of its paid spend and has improved ROAS, though others can't isolate the sales impact. It's a small sample, so treat it as direction rather than proof.
Our read: this is the follow-on to the evidence-over-polish shift we wrote about last week. Shoppers are not only asking "is this good?" but "is this real?"
What to do before Black Friday
Start with a rule, then run tests.
Set a disclosure rule. Keep a simple log of where AI touched each ad: concept, imagery, voiceover, edit. If AI was used in a way a viewer would care about, say so in the caption or the end card. And never stage "process" footage. If the making-of is a recreation, it isn't proof, and the Dior case shows what happens when people suspect it.
Test 1: a "made by hand" cut. Take one hero product and film real hands making, packing or sketching it, for 15 to 20 seconds, labelled plainly as human-made. Run it against your current polished product video on matched budgets. Judge on hold rate, comment sentiment and purchases, not likes.
Test 2: a founder explainer. A candid, claim-backed "what's actually in it" video for one hero product, addressing the dupe question without naming competitors. Naming them carries risk: Beauty Independent reports creators were suspended after trademark complaints over fragrance comparison videos on TikTok, and it cites Circana data that 51% of beauty buyers have bought a dupe. Track repeat purchase alongside first-order ROAS, since founder content is hard to isolate.
Test 3: a dated drop against a discount. Craft pairs well with scarcity. Coach is pulling back from heavy Black Friday discounting in favour of limited colourways, and a Rhode collaboration reportedly drew a 200,000-person waitlist and sold out in about an hour, per e.l.f.'s CEO. Run a teaser, countdown and 24-hour "last colourway" sequence for one limited bundle against a sitewide percentage-off ad, with a holdout, and compare on margin and conversion rate.
One caution on all three: with Black Friday close, keep the test count low. Two clean comparisons beat five muddy ones.
Where we'd be careful
Craft is not a magic word. A hand-drawn aesthetic on an ad that overclaims is still an ad that overclaims, and shoppers are getting quicker at spotting the gap. The evidence behind this trend is also thin: most of the stories above rest on a single report, the founder survey is small, and the Rhode figures come from a CEO's comments rather than independent data. So we'd test the mechanism, visible and verifiable human effort, rather than expect any headline result to repeat. And we wouldn't swing the whole budget behind it. Standard product video still does the job of showing the product clearly, so keep it in the mix as the control and let the numbers decide how much craft earns.
What we're watching
Three things. Meta's AI ad tools reportedly showed discounts advertisers hadn't approved when the offers setting is on in Commerce Manager, so audit yours before the peak. TikTok Shop's BFCM campaign runs 11 to 30 November, and TikTok is pitching the post-Christmas "Q5" window as cheaper inventory. And planned owned-site selling fell from 69% to 46% in a small Modern Retail survey, so your human-made creative needs to work inside shoppable surfaces, not only on your own site.
If you're planning holiday creative and want a sharper test structure or a second opinion on how to disclose AI use, get in touch with us at VHS Digital. We'd be glad to talk it through.
