The Michelin Guide is one of the world's most prestigious restaurant rating systems, but a new investigation by D Magazine has put the guide under a different kind of spotlight, suggesting that artificial intelligence may have been used to write descriptions of several Dallas-area restaurants.
The investigation began earlier this month after a local chef contacted D Magazine dining critic Brian Reinhart to ask whether a recently rewritten description of Sushi Kozy had been generated by AI. Reinhart ran the description through Pangram, an AI detection tool, which flagged the entire text as AI-generated. When Michelin released its updated Texas Guide a week later, he expanded his investigation to include five newly added Dallas-area restaurants.
When asked for comment, a spokesperson told D Magazine that the Michelin inspectors develop the descriptions based on their firsthand dining experiences and that the Guide has its own editorial team, but denied using AI to write or edit its restaurant descriptions.
After running all six restaurant reviews through Pangram, Reinhart found that about 80% were flagged as potentially AI-generated. He then contacted two Carnegie Mellon University professors who co-authored a 2025 paper examining how large language models differ from human writers and their analysis identified grammatical patterns consistent with OpenAI's GPT-4o model, although they could not confirm definitely how the text was produced.
While the investigation raises concerns about how Michelin presents its restaurant recommendations, no evidence suggests inspectors skipped restaurant visits or that AI influenced which establishments received Michelin distinctions.
Check back for more on this developing story.
This article originally appeared on USA TODAY: Did Michelin use AI to write Dallas restaurant reviews?
