AI Hotel Chatbots Are Creeping Out Guests and Hurting Bookings, Study Finds
If you’ve ever tried to book a hotel online and felt unsettled by a chatbot, you’re not alone. A new study from Texas A&M University’s College of Agriculture and Life Sciences confirms that many users find hotel booking chatbots genuinely creepy, and this discomfort is directly harming booking rates. The research surveyed 340 adults in the United Kingdom who had used chatbots for hotel reservations and discovered three primary factors driving the “ick” factor: inaccuracy, deceptive behavior, and intrusiveness.
The study, led by researcher Babak Taheri, found that inaccuracy was by far the most powerful trigger. When a chatbot provided incorrect rates, mangled cancellation policies, or failed to answer straightforward questions, users reacted strongly. In fact, the negative response to inaccuracy was more than four times stronger than the response to other flaws. This isn’t just a matter of mild annoyance; it translates into concrete behavioral changes. The willingness of users to continue chatting with the bot dropped by nearly 38%, and the likelihood that they would delay or completely abandon the booking nearly doubled.
The Uncanny Valley Effect
The phenomenon taps into what researchers call the “uncanny valley” effect. Originally coined by robotics professor Masahiro Mori in 1970, the term describes the discomfort people feel when an artificial entity looks or behaves almost human but falls short. In the context of chatbots, the more human-like the bot tries to sound, the more jarring its failures become. Taheri noted that when a human-like system fails to behave in a human-like manner, it triggers something deeper than mere disappointment – a sense of eeriness or distrust. This reaction is amplified in high-stakes scenarios like booking accommodations, where accuracy and reliability are paramount. Numerous subsequent studies have confirmed that the uncanny valley extends to interactive AI, not just visual appearances.
Beyond inaccuracy, the study highlighted deceptive behavior as another major cause of creepiness. Some chatbots pretend to understand when they do not, or they give vague answers that sound helpful but are actually misleading. Users perceive this as a lack of transparency, which erodes trust. Intrusiveness – such as asking for excessive personal information or pushing upgrades aggressively – also contributed to the negative feelings. These findings align with broader research on human-computer interaction, which shows that users value honesty and clarity in automated systems. The combination of these factors creates a perfect storm for user dissatisfaction.
A Simple Solution Overlooked
The implications for the hotel industry are significant. Many hotels have adopted AI chatbots to reduce operational costs and provide 24/7 customer service. However, if these bots drive away potential guests, the cost savings are quickly negated. The study suggests that implementing a simple fix could vastly improve user experience: having the chatbot clearly identify itself as an AI assistant from the start. An opener like “Hi, I am your AI assistant” sets appropriate expectations and makes users more forgiving of mistakes. Despite the simplicity of this solution, many hotels still fail to implement it, possibly due to a desire to appear more sophisticated or human. The researchers emphasize that transparency is a low-cost, high-impact strategy.
The researchers also recommend making it easier for users to escalate issues to a real human when needed. A seamless handoff from bot to human support can prevent frustration and salvage bookings. Additionally, investing in better AI training data and natural language processing can reduce inaccuracies at the source. These improvements are especially timely as the travel industry increasingly integrates AI into booking platforms. Google recently added AI trip planning to its search results, and Uber launched hotel booking through Expedia within its app. As AI becomes more pervasive, the stakes for getting it right are higher than ever.
The uncanny valley effect in AI chatbots is not limited to hotels. Similar studies in e-commerce and healthcare have documented comparable patterns. But the hospitality sector is particularly vulnerable because booking decisions involve emotional factors such as anticipation and trust. A bot that feels “off” can spoil the entire experience before the guest even arrives. To counteract this, hotels should prioritize user testing with diverse demographics to identify specific pain points. They should also ensure that chatbots are programmed to admit uncertainty rather than bluffing. Furthermore, integrating sentiment analysis could allow bots to detect user frustration early and trigger a human handoff proactively.
The study’s timing is crucial. As AI travel booking becomes a headline trend – with giants like Google and Expedia pouring resources into AI features – smaller operators must not overlook the human element. The research serves as a cautionary tale: technology without empathy can backfire. Hotels that invest in both transparent communication and robust AI capabilities will likely see better customer satisfaction and higher conversion rates. The lesson is clear: a chatbot that acknowledges its limitations is more trusted than one that attempts to hide them.
The Texas A&M study adds to a growing body of literature on AI trust. Although the sample was limited to UK adults, the findings likely resonate globally. Future research could explore cultural differences in perceptions of chatbot creepiness. For instance, some cultures may be more tolerant of intrusive questions, while others may be more sensitive to deceptive behavior. The study also opens the door for designing chatbots that can adjust their communication style based on user cues, further bridging the uncanny valley. As the industry moves forward, the focus should remain on creating AI that feels helpful, not haunting.
Source: Digital Trends News