Tripadvisor AI Summaries Downplay 3 Hotels' Safety Complaints, Which? Finds
Updated
Updated · The Guardian · Jul 1
Tripadvisor AI Summaries Downplay 3 Hotels' Safety Complaints, Which? Finds
2 articles · Updated · The Guardian · Jul 1
Summary
Which? found Tripadvisor’s AI hotel summaries softened severe guest warnings, including a Cape Verde resort it called “spotless” despite complaints of food poisoning, raw chicken, flies and mice.
The investigation also flagged a Dominican Republic hotel where guests said mains water failed and many in a wedding party fell ill, and a Turkey hotel where harassment complaints were reduced to “friendly” service and minor lapses.
Tripadvisor said it is refining the tool and checking mismatched examples, but defended the feature as a quick guide alongside its billion-plus reviews rather than a replacement for reading them in full.
The Cape Verde property’s summary has been removed, while researchers said AI often sanitises sharp criticism by flattening rich consumer feedback into milder sentiment.
The findings add to wider scrutiny of AI summaries after Google earlier removed some health overviews that a Guardian investigation said risked harming users.
As AI sanitizes reviews and silences warnings, how can travelers uncover hidden dangers before booking?
Who is liable when a platform's AI summary misleads a traveler into a dangerous situation?
Tripadvisor AI Summaries Omit 3 in 5 Critical Safety Complaints, Investigation Finds
Overview
A recent investigation by the consumer group Which? uncovered that Tripadvisor’s new AI-generated hotel review summaries often downplay or omit serious safety complaints, putting travelers at risk. This problem arises because the AI, trained mostly on neutral or bland data, tends to 'sanitize' sharp criticisms, making negative experiences seem less severe. As a result, important safety issues can be hidden from travelers who rely on these summaries. The findings have sparked debate about the reliability of AI in sharing critical information and highlight the urgent need for better oversight and transparency in AI-powered review systems.