James Rowe: A Which? investigation has cast doubt on the accuracy of Tripadvisor’s AI summaries. Here’s what you need to know.
Hello, I’m James Rowe and welcome to Which? Shorts - your free weekly insight into Which? magazine, as well as our Money, Tech, Travel and Gardening titles too.
Today I’m bringing you a piece that our travel expert Kate Pasola wrote for the July/August issue of Which? Travel, all about Tripadvisor and its AI summaries on hotels.
AI summaries are becoming increasingly prevalent, from Google searches to news articles, they’re designed to give us a quick overview of the topic at hand. But our investigation into Tripadvisor has shown that the information isn’t an accurate reflection of the reviews left by real travellers.
Here, we shine a light on Tripadvisor’s AI summaries, with some stark examples of just how different the summaries are compared to reality.
Here’s Kate’s piece, adapted for the podcast, this week read by Erica McKoy.
—
Erica McKoy: Trust Tripadvisor’s words alone and you’d expect bliss at the Riu Palace Santa Maria hotel in Cape Verde. This five-star all-inclusive resort is, it says, a beachfront resort ‘popular with many travellers’. ‘Spacious rooms’, ‘lively entertainment’, and ‘friendly staff’. It also highlights the ‘diverse restaurants’ which, it says, ‘earn rave reviews’.
Those words are directly from the Tripadvisor ‘reviews summary’ at the top of the page and created by AI. Scroll down to the real reviews and it’s a different story.
Recent guests at the Riu Palace reported ‘exceptionally poor hygiene,’ ‘no basic cleaning or hygiene standards’, and that the food was ‘awful, bland, unsafe and inedible’.
One guest said that she was served raw chicken. Another review shared photographs of flies and birds in the buffet food, and another spotted ‘dead little roasted mice by the sitting area’ on her ‘nightmare’ holiday. One guest whose whole family fell ill wrote: ‘This place will destroy holidays, and has the potential to take lives.’
When we checked in March, there were a staggering 102 mentions of food poisoning at the Riu Palace. That doesn’t necessarily mean 102 guests suffered from food poisoning, but many clearly did.
The resort also had 32 one and two-star reviews posted between December 2025 and April 2026 alone, 14 of which say at least one member of the party fell seriously ill with some form of food poisoning.
Many were hospitalised, some flew home early. One guest died this year.
A spokesperson from RIU Hotels & Resorts told us its hotels in Cape Verde ‘follow the strictest international health and hygiene standards, certified by external prestigious consultancy firms, specialized in health and safety’.
However, the hotel is now involved in a group legal action representing at least 412 holidaymakers who say they became ill after staying at the property, with seven deaths reported since 2023.
None of this is Tripadvisor’s fault. But when reading the AI summary of the resort, provided by the world’s most popular hotel review platform, you would have no idea of the dangers potentially posed by staying there. Instead, Tripadvisor surmises the hotel’s cleanliness as ‘spotless’.
Another of Tripadvisor’s AI tools, an interactive trip planning bot called Ollie, also failed to warn us about poor hygiene. When asked directly about the risk of contracting food poisoning at the Riu Palace, Ollie told us food poisoning was ‘quite unlikely’, and that the resort had a ‘strong reputation for high hygiene standards’.
When we asked Tripadvisor about its AI summaries, it said it prioritises ‘transparency and impartiality’ and its summaries ‘surface a range of both positive and negative community feedback associated with listings’.
Tripadvisor told us its AI chat assistant tool Ollie ‘draws from a selection of reviews based on detail and recency, and matches by language and context’, but it added that it is a ‘product in development’, and that it is now actively looking into several examples we provided where reviews did not match the intended property.
There are other dangers Tripadvisor’s summaries don’t share.
At Kaia Coracesium on the Antalya coast, several reviewers who visited last summer wrote they felt unsafe due to repeated sexual harassment from male hotel staff, including inappropriate jokes and gestures, and repeated requests to connect on social media.
Two different guests reported that a male member of staff followed their daughters to request their social media details. In one of these cases, the reviewer says a restaurant worker ‘followed her up the stairs to our room – obviously, this scared her and she didn’t feel comfortable walking around on her own’.
The Tripadvisor AI review summarises the service as ‘friendly’.
The closest it comes to referring to these serious allegations is: ‘Lapses in service noted by a few’. This, at least, demonstrates that its AI is aware of these allegations and reasons that the opinion of ‘a few’ could be worth mentioning. So why isn’t it doing so consistently – and when it does, why does it downplay the severity? Repeated accusations of sexual harassment by staff is not a lapse.
Many travellers would almost certainly consider an alternative hotel based on these ‘red flag’ reviews. According to The Safer Tourism Foundation, nearly two thirds of women prioritise safety when choosing accommodation. This sort of information is vital to their planning.
Which? reached out to all hotels mentioned, but none apart from Riu responded with comment.
Tripadvisor said its summaries ‘use large language models and natural language processing to read recent reviews, identify the most common themes and then turn those themes into short, plain‑English overviews’, and that its goal was to ‘make content from Tripadvisor’s reviews and opinions as easy as possible to digest, but also to capture and highlight the broad spectrum of positive and negative opinion without favouring one sentiment or the other’.
It told us summaries are updated on a monthly basis ‘and are rooted in the previous 12 months of reviews at the time of each update’, and that reviews are treated equally regardless of rating, highlighting what reviewers mention most often.
It added: ‘The system does not deliberately up-weight older or more positive ones, or suppress specific comments… As a result, issues that are serious but relatively rare may not be identified as a dominant theme, even though they are clearly important to the travellers who experienced them. Conversely, things that come up in a large share of reviews - for example, “great location”, “busy pool” - tend to appear prominently in summaries.’
This suggests that phrases like ‘I was hospitalised’ or ‘I was followed to my room by a member of staff’ are treated with the same importance as ‘I had a relaxing stay’ by Tripadvisor.
We asked Tripadvisor to confirm whether this was in fact the case. It didn’t clarify, but told us its systems ‘automatically suppress AI summaries for listings that feature warnings from travellers about serious safety incidents such as death, drugging or sexual assault’.
One thing is certain – Tripadvisor’s roll-out of some AI tools has made it money.
Head of data and AI, Rahul Todkar, said that ‘users engaging with our AI-powered features, like trip planning, are showing two to three times more revenue’.
Tripadvisor has since clarified that this revenue uplift was only through a separate itinerary planning tool ‘during its beta testing phase’, and that it has ‘no connection to content integrity’.
But the point remains, AI tools, no matter how lucrative, should only be rolled out when they’re truly safe to use – and Tripadvisor’s aren’t.
