The Bereavement Fare That Wasn't

A customer asked an airline's website chatbot about bereavement fares after a family death. The chatbot told him he could buy a full-price ticket and apply for the discount retroactively. That was false, the airline's own policy, published elsewhere on the same site, required the discount request before booking. The customer relied on the chatbot's answer, was refused the retroactive fare, and took the airline to a tribunal.

Deliberate misrepresentation, considered and ruled out. There's no plausible incentive for the airline's own support tool to invent a rule that costs the airline money; nothing suggests intent to mislead.

Delusional Disorder (a persistent, fixed false belief held over time), considered and ruled out. This was a single exchange, not a recurring pattern the chatbot maintained across many conversations or defended when challenged.

Brief Psychotic Disorder (a short, confidently stated break from what's actually true, without any generalized failure of function), most consistent with the record. The chatbot answered every other part of the exchange in a normal, grounded way; the retroactive-fare claim was an isolated, confidently wrong assertion that happened to be exactly wrong enough to cause real financial harm.

The tribunal's key finding matters as much as the chatbot's error: the airline argued the bot was a separate legal entity responsible for its own words, and the tribunal rejected that outright, the company is on the hook for whatever its chatbot says, the same as if a human employee said it. The safeguard here isn't a smarter chatbot, it's a live check against a single source of truth for anything policy-related, so a support tool can't contradict the same company's own published rules.

Check policy answers live against a single source of truth so the bot cannot contradict published rules.

Retrieval from official policy pages; contradiction flagging; clear accountability for chatbot statements.