A swing and a Mich for artificial intelligence?
The Michelin Guide, revered as the world’s foremost restaurant rating system, The Michelin Guide, revered as the world’s foremost restaurant rating system, tasked AI to write its 2026 review of Dallas, Texas eateries, claims a jaw-dropping new report.
And it’s leaving a foul taste in foodies’ mouths, with outraged folks across the internet saying the offense has “burned” the iconic trademark’s reputation “to the ground,” and others calling its star-spangled guide “a fart in the wind.”
Although bots à la ChatGPT and Claude are able to do a lot — from generating testimonies in murder trials to launching phony culture wars — the platforms, apparently, cannot fool the suspicious chefs calling out Michelin for allegedly using a large language model (LLM) to scribe its Dallas-area guide, per a D Magazine investigation.
The analysis, “including consultation with detection software and academic experts,” reveals that most or all of Michelin’s new Dallas entries were written by LLM software, according to the outlet, initially tipped off by a concerned local cook.
Representatives for Michelin did not immediately respond to The Post’s request for a comment regarding the allegation.
The company did, however, deny the use of AI in its Dallas review to D Magazine, saying, in part, “All Inspector notes appearing on the MICHELIN Guide website and app are developed collaboratively by the Inspection team based on their firsthand dining experiences.”
“AI is not used in the writing of these reviews.”
But an overabundance of flowery descriptions, like the “serene” dining room at Mābo and the “carefully sourced” ingredients at Sushi Kozy, plus numerous grammatical anomalies, triggered alarms for AI detector Pangram, as well as Carnegie Mellon University professors Alex Reinhart and David Brown, the findings revealed.
Pangram immediately flagged Michelin’s report as “100%” AI, with Reinhart and Brown pointing to GPT-4o — an outmoded, notoriously saccharine LLM — as the likely AI author.
The study also compared the review to past Michelin appraisals throughout the city, as well as other major cities worldwide. It found that AI had not been used in any guide other than this year’s Dallas assessment.
Reinhart and Brown speculated that a Michelin inspector — one in an army of anonymous culinary experts who score meals based on a “stars” system — fed the chatbot a few notes, then sprinkled their own thoughts about each Dallas hotspot into the write-up.
“There are chunks of it that seem more AI-ish than others,” said Brown. “Afterwards they might have changed, ‘Oh, this looks good, but I don’t like this sentence.’”
The D Magazine study highlights Michelin’s “most perverse” sentence in its review of Resident Taqueria, which reads, “Thoughtful combinations, bold flavors and generous portions reflect the kitchen’s creativity,” noting the line’s seemingly nonsensical, disingenuous vibes.
The accusation against Michelin rises amid trending AI controversies, including President Trump’s wild tiger-taming clip, and a string of lawsuits arguing that ChatGPT grooms children for self-harm.
But it’s certainly not the first time a trusted brand has been fingered for allegedly faking the funk.
McDonald’s is currently embroiled in a federal court case, per claims it violated US antitrust laws by conspiring with franchisees to fix menu prices through an AI-powered system.
California café, Grind and Unwind, recently caught viral “false advertising” flak after plastering “AI-slop” menus in its windows, spurring an avalanche of online criticism, negative reviews and vandalism.
Brown openly questioned Michelin for potentially jeopardizing its decades-long relationship with chowhounds who laud the guide as a reliable resource.
“Food writing is so human,” the research said in the study, in part. “There’s something very human about it. Getting AI to generate it seems a curious choice.”
“Why would you do that?”
Read the full article here
