Tripadvisor AI tools creating misleading summaries, finds Which?

Travel platform Tripadvisor is using AI to generate misleading hotel review summaries that do not disclose serious safety concerns, including food poisoning, poor hygiene, and allegations of sexual harassment, according to Which?.

An investigation by the consumer watchdog found that the platform’s AI-generated review summaries presented positive descriptions of several hotels despite numerous recent guest reviews detailing significant health and safety issues.

Which? said Tripadvisor’s new AI summaries, which appear at the top of hotel listings, could giving travellers an inaccurate impression before they read individual customer reviews.

One example highlighted by Which? is the Riu Palace Santa Maria, where the AI summary describes the resort as offering “spotless” cleanliness and restaurants that receive “rave reviews”. However, Which? Said recent reviews on the platform include complaints about poor hygiene, raw food, contamination, and widespread reports of food poisoning.

According to the investigation, there were 102 mentions of food poisoning linked to the resort when Which? reviewed the listing in March, while multiple guests reported being hospitalised. The hotel is also the subject of legal action involving more than 400 holidaymakers who say they became ill after staying there.

Which? also tested Tripadvisor’s AI travel assistant, Ollie, by asking about the risk of food poisoning at the hotel. The chatbot replied that food poisoning was “quite unlikely” and claimed the resort had a “strong reputation for high hygiene standards”.

The consumer group found similar issues at several other properties, including hotels in Mexico and the Dominican Republic, where AI-generated summaries focused on positive aspects despite recent reviews describing illness, sewage smells, mould, and poor maintenance.

Friendly service

Which? also highlighted a hotel in Turkey where guests had reported repeated sexual harassment by staff, while Tripadvisor’s AI summary simply described the service as “friendly”.

For comparison, Which? said Google’s AI-generated hotel overviews more clearly flagged health concerns, including warnings about illness outbreaks and food safety issues at the same resort.

Rory Boland, editor of Which? Travel, said the platform has a responsibility to investigate the accuracy of its AI tools and urged users to scroll past these summaries and look at guest reviews.

“Tripadvisor may insist users can still fact-check its summaries against real reviews, but this ignores the fact that it made the decision to push these summaries to the very top of the page,” he added. “This failure to surface critical safety information is unacceptable and potentially life-threatening.”

Tripadvisor disagreed with the investigation’s allegations and said that its AI summaries have been designed to provide snapshots based on high volumes of user generated content and explicitly are not intended to replace individual reviews.

The company added that users can easily click to see the traveller quotes behind each review element or access all reviews for that listing, eliminating any need to blindly trust AI-generated content.

“We also have comprehensive safeguards in place to ensure important safety information is properly reflected across our platform” the spokesperson added. “Our AI systems are designed to capture all types of traveller feedback and we continuously monitor and refine our models.

“Our systems automatically suppress AI Summaries for listings that feature warnings from travellers about serious safety incidents such as death, drugging or sexual assault, helping ensure this content is highly visible to our community.

“Our AI-powered chat assistant draws from a selection of reviews based on detail and recency, and matches by language and context. We are actively looking into the examples where reviews did not match the intended property, and, as with all products in development, we will use the feedback to improve the traveller experience.”

The spokesperson emphasised that no review content has been suppressed or hidden by the introduction of the tools and said the suggestion they pose danger to travellers is an unfounded claim that seems designed to generate controversy, rather than inform readers.

“We believe our community understands that AI technology is still developing and has the common sense to check any AI advice against Tripadvisor’s billion-plus reviews and contributions.”