A recent trip to Rome left me impressed with my dinner reservations, thanks to AI-powered recommendations from GPT-5. I had asked the model for suggestions near our hotel and was directed to Babette, a short walk down Via Margutta. The meal was one of the best I’ve ever had.
Curious about how the model chose this restaurant, I asked it to explain its decision-making process. While the response was complex, several factors contributed: rave reviews from locals, notices in food blogs and Italian press outlets, and a unique blend of traditional Roman and contemporary cooking. The proximity to our hotel didn’t hurt either.
However, not everyone is convinced about GPT-5’s recommendations. Some commenters question the model’s accuracy and reliability, citing instances where responses seem “confabulations loosely tethered to reality.” With concerns like these, it raises questions about the future of AI-powered services in the travel industry.
Source: https://www.wired.com/story/can-ai-escape-enshittification-trap