AI-powered chatbots are becoming increasingly common on hotel booking websites, helping travelers search for rooms, compare prices, and answer questions.
But new research suggests these virtual assistants may also be making many customers uncomfortable—and in some cases, driving them away before they complete a booking.
The study, published in the International Journal of Hospitality Management, found that travelers often feel uneasy when hotel chatbots appear inaccurate, unreliable, overly intrusive, or misleading.
Researchers say this feeling of “creepiness” can reduce trust, cause frustration, and even lead customers to abandon their reservations altogether.
The research was led by scientists at Texas A&M University, who surveyed 340 adults in the United Kingdom who had used chatbots while booking hotels online.
The researchers discovered that inaccurate information caused the strongest negative reaction. When chatbots gave incorrect room prices, misunderstood cancellation policies, or failed to answer questions properly, users became far more likely to stop interacting with the system.
According to the study, these uncomfortable feelings reduced people’s willingness to continue chatting with the AI by nearly 38%. The discomfort also nearly doubled the chances that users would delay or completely abandon their booking decisions.
Researchers explained that the problem becomes even more noticeable when chatbots try to sound very human. This relates to a psychological phenomenon known as the “uncanny valley,” where technology that behaves almost like a person—but not quite perfectly—can feel unsettling rather than helpful.
People naturally expect a human-like chatbot to communicate smoothly and accurately. When it fails, the mismatch between expectations and reality creates a stronger emotional reaction than a simple technical mistake.
Lead researcher Babak Taheri said users often feel threatened or manipulated when a chatbot behaves inconsistently or seems deceptive.
However, the study also found that companies can reduce some of this discomfort through something surprisingly simple: honesty.
When chatbots clearly introduced themselves as AI at the beginning of a conversation, users became more forgiving of mistakes. Instead of assuming the system was intentionally misleading them, people were more likely to see the errors as normal AI limitations.
For example, researchers suggested that hotel websites begin conversations with a short message such as, “Hi, I’m your AI assistant.”
At the same time, transparency was not a perfect solution. When users already knew they were talking to AI, they became more sensitive to anything that felt deceptive or unnatural. This means companies still need to improve chatbot quality and accuracy, not just announce that AI is being used.
The researchers recommend that hotel and travel companies invest in better AI systems that can more reliably answer important questions about room availability, pricing, refunds, and cancellations. They also suggest giving customers an easy option to switch from the chatbot to a real human employee whenever needed.
The study highlights an important lesson for businesses adopting AI customer service tools: convenience alone is not enough. Travelers also want transparency, accuracy, and trust—especially when spending money on important travel plans.
