AI's false information is weaponizing customer entitlement in service work

When a server informs a diner that a dish contains shellfish and the customer can't be moved because ChatGPT told them otherwise, a simple customer-service friction has become a potential safety hazard. This dynamic is playing out across hospitality and retail workplaces: generative AI is amplifying customer obstinacy, and the fallout lands on frontline workers.

The allergy problem at the core#

Madison, a server at a Michelin-recommended New York restaurant, has noticed alarming shifts in how guests respond to critical dietary information. Customers dismiss her expertise on ingredients, insisting that an AI chatbot's assurance overrides her knowledge—even when safety is on the line. AI models routinely fabricate details: fictional wine brands, nonexistent menu items, dishes the restaurant has never served. When Madison tries to clarify, patrons treat the correction as pushback, not as fact-checking.

The sommelier role has been particularly disrupted. Guests now ask ChatGPT instead of trained staff which wines to pair with their meal, sometimes requesting bottles that either don't exist or aren't in stock. Madison and her colleagues have adopted a gentler framing to soften these moments—describing the AI's error as a quirk of their restaurant's specialty list rather than spelling out the truth: the AI simply made it up.

Hallucinations spreading beyond restaurants#

The problem extends far beyond wine service. Movie theater workers, baristas, and Apple Store employees all report fielding requests for products and features that don't exist, based on AI output. On Reddit and social media, customer-service staff are openly discussing how chatbot use has degraded their jobs.

Jason, a river ranger on federal lands in the Mountain West, has watched visitors arrive with AI-generated trip itineraries containing fatal errors: campsites that don't exist, routes that would require paddling upstream. Even when a ranger points to a map and explains the mistake, some travelers remain skeptical, trusting the AI over expert guidance and printed resources.

Eroding professional expertise#

Casey, an after-school program teacher in New York, has observed parents turning to ChatGPT for childhood development questions—and then overriding her assessments. One family consulted an AI tool to design their child's sleep schedule, leading to insufficient rest. When Casey noted the child needed several more hours nightly, the parents held firm: their chatbot disagreed.