Hotel Forgets to Delete Prompt, Exposing LLM-Written Responses to Guest Reviews
abhishekcode42 · x · 2026-10-02
A guest caught a hotel accidentally pasting its LLM prompt into a review reply. The leaked instruction: acknowledge the guest felt the property is "not suitable for leisure travellers" rather than defending the location — fresh evidence that hospitality brands are quietly using LLMs to mass-produce customer-service replies.
More from Fun
- 58 Years Ago Kubrick Sketched the AI Alignment Problem: HAL Was Just Given Conflicting Goals — Hesamation · 2026-10-02
- ChatGPT voice makes a shopping list infographic her husband can't mess up — VoidStateKate · 2026-10-02
- One prompt, one song: Opus 5.5 runs 12 hours autonomously and delivers a finished piece — minchoi · 2026-10-02
- Director crafts surreal AI short film 'sonder' with MiniMax H3 — Hailuo_AI · 2026-10-02
- 'GLM 5.3 and Kimi K3 topping this list is a national security issue,' quips US tech exec — matt_slotnick · 2026-10-02
- Gary Marcus seeks controls on viral 'LLM pain' paper, says author disclaims pain claim — GaryMarcus · 2026-10-02