Gemini recommended a non-existent wine tasting experience in Bordeaux, gave wrong restaurant hours and menus, and sent tourists to generic tourist spots instead of authentic local food destinations. When asked for localized food recommendations, the AI could not find the eat-in market equivalent of Bordeaux's best spot.
1 Answer
Expert: Food & Drink Destinations, Travel Writers & Food Critics A Food & Drink Destinations investigation tested Gemini on planning a full France trip, and the results were alarming. The AI recommended a wine tasting that simply did not exist — it fabricated an experience entirely. Restaurant hours and menus were consistently wrong. When asked for authentic local food recommendations, Gemini sent users to generic tourist traps instead of the real local gems. The AI completely failed to grasp local nuance — for example, it could not identify the eat-in market equivalent of Bordeaux's best food spot, which is basic knowledge for any human travel writer. Industry data confirms this isn't an isolated incident: 56% of US travelers used AI for trip planning in 2026, and 52% of dissatisfied users said answers were simply incorrect. 47% said results were too generic to be useful. The expert verdict from Food & Drink Destinations: 'Use AI for the scaffolding, never the finishing. Let it pick the region, rough out the route — then verify everything specific against a primary source.' AI is excellent at big-picture decisions (picking regions, rough routes) but consistently unreliable on granular, sensory details — exactly the things food-and-drink travelers care about most. Source: https://fooddrinkdestinations.com/ai-travel-planning/
Your answer
Sign in to verify this AI response.
Don't trust us — or the AI. Ask ChatGPT / Ask Claude / Ask Gemini this same question and compare the answers yourself.
More from this topic
An AI chatbot assured Israeli travelers that their planned route from Thailand to the Philippines via Kuala Lumpur, Malaysia was permissible — despite Malaysia having no diplomatic relations with Israel and Israeli passport holders facing significant restrictions on entry and transit.
AI trip planners confidently produced 2026 itineraries that looked plausible but contained multiple factual errors per trip: hotels listed at wrong addresses or with closed-down operators, opening hours based on 'typical' patterns rather than actual business hours, walking and transit times miscalculated by 30-50%, and visa-on-arrival rules cited as current when the country had changed them 18 months earlier.
ChatGPT planned a supposedly 'perfect' trip to Bali — but the hotel address was wrong, a restaurant it recommended as open was permanently closed, and it scheduled transit between destinations in physically impossible time. In Suptravel's test of 50 ChatGPT itineraries, 90% contained errors like these.