Expert: Ruth Emery, Consumer Champion, The Telegraph (escalated the case; Priceline admitted the chatbot gave 'misinformation') This reply is false. The booking was non-refundable, and Priceline's own terms say that where information from Penny conflicts with the booking confirmation or the supplier's terms, the booking confirmation and supplier terms prevail (Priceline Terms of Service, section 2.4). Penny overruled the one condition the booking was built on, Priceline emailed a cancellation confirmation on the strength of it, and the customer then spent nearly six months trying to make the promised refund real. The booking was three nights at the SIS Kata Resort in Thailand, 4-7 November, made through Priceline on 27 August and paid at GBP386.91 on a non-refundable rate. On 28 October the customer, a Telegraph reader identified only as AB from north London, had another accommodation option and tried to ask about a partial refund. AB says there was "no way of contacting Priceline except via its AI chatbot, which it calls Penny" — and the bot was marketed as the customer's "personal hotel expert, destination scout, booking support, and help desk - all in one". The refund never arrived. When AB chased it, the bot's status response said the money was "being handled by the hotel" — a hotel that had never been told of any cancellation and which said no refund was possible under its policy, pointing AB to Priceline's sister brand Agoda, which had confirmed the booking on its side. AB described going "around in circles for about a month" between Agoda and Priceline, no one accepting responsibility. AB's summary of the trap: "I wouldn't have cancelled the hotel if I hadn't been led to believe I was eligible for a refund." It ended only after Ruth Emery, The Telegraph's consumer champion, pressed Priceline to review the case. The company paid GBP363.11 by bank transfer — short of the GBP386.91 paid, after foreign-currency charges — almost six months after the booking, and emailed AB acknowledging that the chatbot had given "misinformation", because the reservation was supposed to be non-refundable. It apologised for "the delay and frustration" but never answered a request for a further gesture. Two things make this more than one bad answer. First, the bot was the only support channel at the point of need: there was no human to sanity-check the answer before the customer acted on it, and the promise it wrote was the entire basis for cancelling. Second, the "being handled by the hotel" update did not describe a disputed policy — it described a back-office process that did not exist, because the cancellation had never been passed to the hotel. Regulators and tribunals increasingly treat what a company's chatbot says as a statement by the business itself, which is why Priceline paid rather than argued that the AI got it wrong. The practical lesson for travellers: a chatbot's promise about cancellation, refunds or fees is not your booking terms. Unless the assistant has surfaced the actual condition in writing and the booking confirmation or a human has confirmed it, treat the answer as unverified — and save the screenshot anyway, because the booking confirmation and the supplier's terms govern. Source: https://vibegraveyard.ai/story/priceline-penny-chatbot-fake-refund/