Tripadvisor AI Summaries Fail to Warn of Food Poisoning, Sexual Harassment at Hotels

Tripadvisor's new AI-powered review summaries are masking serious safety issues at hotels, according to a Which? investigation. The summaries, placed at the top of hotel pages, condense guest feedback but omit reports of food poisoning, sexual harassment, and hygiene failures. The watchdog found several examples where AI summaries gave glowing recommendations while recent reviews told a very different story.
Key Findings
- Riu Palace Santa Maria (Cape Verde): AI summary calls it 'spotless' with 'rave reviews' for dining. Yet 102 mentions of food poisoning were found, with 32 one- and two-star reviews between Dec 2025 and Apr 2026. 14 mention serious illness, multiple hospitalizations, and one death. The hotel faces a group legal action representing at least 412 holidaymakers, with seven deaths since 2023.
- Garza Blanca (Cancun): AI summary praises 'immaculate cleanliness'. Recent reviews report multiple guests falling ill, including a wedding party.
- Occidental Caribe (Dominican Republic): AI summary notes 'abundant' amenities and only 'inconsistent' cleanliness. Reviews describe sewage smells, mold, no running water (guests showered with bottled water), and half a 68-person wedding party falling ill.
- Kaia Coracesium (Turkey): AI summary describes sexual harassment from staff as 'lapses in service'. Reviewers reported repeated harassment, including staff following daughters to request contact.
Tripadvisor's interactive planning bot, Ollie, also failed. When asked about food poisoning risk at Riu Palace, Ollie replied it was 'quite unlikely' and the resort had a 'strong reputation for high hygiene standards'.
Tripadvisor responded that summaries 'surface a range of both positive and negative community feedback' and that Ollie 'draws from a selection of reviews based on detail and recency'. It said it is actively investigating the examples provided by Which?.
Which? checked in March 2026 and found 102 mentions of food poisoning at Riu Palace alone. The AI summaries are generated from a subset of reviews, but the criteria for selection appear to exclude the most critical ones.
📖 Read the full source: HN LLM Tools
👀 See Also

Claude outperforms Gemini, ChatGPT, and Grok in real-time Python coding challenge
A developer tested Claude, Gemini, ChatGPT, and Grok in a real-time Python coding tournament where AI-generated bots competed to find words on a 15×15 letter grid. Claude won decisively.

AI Hyperscaler Debt Surges 974% to $225B in 2026, Hidden Debt Hits $1.65T
Hyperscaler bond issuance jumped 974% to $225B in H1 2026, with hidden debt at $1.65T. S&P flags investor fatigue as premiums rise.

Diagnosing Operational Drift and Task Amnesia in OpenClaw with Gemini 2.5 Flash on Proxmox
OpenClaw users report issues with persistent workflows on a Proxmox VM, citing operational drift and task amnesia. Despite stable performance in one-off tasks, the Gemini 2.5 Flash model struggles with automation and memory in this setup.

Anthropic Policy Update Bans Third-Party Tools for Claude Pro/Max Users
Anthropic updated their policy in February 2026 to explicitly ban any script, wrapper, or third-party tool usage with Claude Pro or Max plans, resulting in lifetime bans for users who violate this policy. High-tier Max plan users engaging in heavy coding sessions are being targeted in a March 2026 enforcement wave.