Will AI Solve the Knowledge Gap on Air Passenger Rights?

Berlin, 20 August 2026 – Nobody Googles anymore. They ask. Nearly half of US adults now use an AI chatbot, roughly double the share from just two years ago - and the single biggest reason they open one is to look something up (Pew Research Center, 2026). In travel specifically, that shift has already crossed the halfway mark: 56% of US leisure travelers now use AI somewhere in their trip planning (Phocuswright, 2026). When a flight gets cancelled or seriously delayed, the question "Can I get compensation for this?" increasingly goes to ChatGPT or Claude before it ever reaches a search bar.

Where Trust Runs Out

It's not only a trend. The travel industry has given people plenty of reason to stop trusting the old way of finding out, in particular when flight disruption occurs. An AirHelp survey shows that significant delays and cancellations leave 68% of EU passengers and 72% of US passengers stressed and frustrated. A third of travelers say they've lost confidence in airlines to handle disruptions at all, and half say they no longer care who fixes the problem, human or AI, only that someone does (Ada/Dynata, April 2026).

Compensation is one way to make up for the mess. But how do you know what rule actually applies to you? It's clearly not going to be the airlines closing that knowledge gap: AirHelp's research found only 25% inform passengers of their rights during a disruption at all. And if you ever tried looking up the law as it applies to your exact case, you'll find it's not curiosity that kills the cat. Complexity does.

So here's an uncomfortable question: if people have already changed where they ask, why hasn't the answer moved with them?

The Law Isn't Missing. It's Invisible.

In the EU Passengers bumped, delayed or cancelled are often owed up to €600 in compensation. Yet AirHelp's own data shows 60% of eligible travelers never pursue the compensation they're owed. Not because they don't deserve it. Because they don't know it exists or don’t understand that it applies to them.

It gets worse: More and more eligible claims become initially rejected - a statistic that rose from 52% in 2025 to 58% in 2026. Most rejected passengers never push back, even when the rejection wouldn't survive a real look under the law. An airline says no, and that's the end of the story. Not because the passenger was wrong, but because nobody told them that the airline was.

Right Answer, Right Address

Since 2016, AirHelp has built AI into every step of an active claim: Documents get checked the second they're uploaded, flight data gets read the moment it settles, airline messages get translated instantly across 19 languages. A traveler's trip history builds itself from their inbox - typos included as the system knows "Rajanair" means Ryanair.

Yet none of it matters to the 60% who never got far enough to file a claim.

So AirHelp is creating answers where questions actually gets asked. Built AI cloud technology provider Alpic, its compensation checker now lives inside ChatGPT and Claude. Type "My flight from Barcelona to Milan was cancelled. Can I get compensation?" in one of eight languages, and get an answer in the same breath: eligible, with a payout estimate and the regulation behind it - or not, with a plain reason why.

Ask ChatGPT or Claude the same question on its own, and it can explain the general rules - but it can't look up your actual flight. It doesn't know if your delay was three hours or five, whether the airline gave proper notice or which of a dozen overlapping regulations applies to your specific route. Instead, AirHelp’s new checker taps into AirHelp's own flight and regulatory data to identify the exact flight, apply the exact rule and calculate the exact compensation amount, turning a plausible-sounding answer into a verified one.

The app is one part of a broader effort: AirHelp continues improving how customers can find relevant information about their rights and eligibility in the place where they're actually looking for it.

Try it yourself

AirHelp is inviting journalists to test the dialogue directly: describe a real disrupted flight to the checker in ChatGPT or Claude and see what comes back.