United Airlines CEO Scott Kirby said AI has proven more capable at interpreting labour contracts than people, speaking at the Bernstein Annual Strategic Decisions Conference.
“One of the interesting things that we've used [AI for] is these labour contracts that have been in existence for decades,” said Kirby. “They have been in existence for decades, and they have provisions that were written at a time — in a different era — that wouldn't apply today. All these contract provisions cross, they confuse each other, and they're hundreds of pages long.
“We have specialists for interpreting the contracts. So in this situation, what does the contract mean and what happens with your schedule, etc. And it turns out AI is better than people, as well as more accurate and obviously faster at getting it done.”
United's flight attendant union said it reached an “industry leading” tentative labour deal for United's roughly 28,000 flight attendants, the union said on May 23, 2025. The flight attendants secured 40% of “total economic improvements” in the first year, including retroactive pay, a signing bonus. Locally elected leaders met on May 29th and 30th to review the full details before it is sent to members to be voted on. The union did not provide further details on the contract.
Kirby said he was “glad” to have got the deal done. “They got an industry-leading deal, which they deserved and I'm happy for them,” he added. “Financially, it's been in our guidance and is still in our guidance. We'll update the full year when we come out the next time, but it is financially in the guidance."
With AI supporting the contract successfully, as well as its current best use case being at its call centres, Kirby said the company is continuing to invest in the technology to find other best use cases for AI.
“We're trying to use it as one of the use cases that everyone thinks ought to be good is predictive maintenance on airplanes,” he continued. “And we have some isolated cases where that's worked, but the truth is it hasn't worked out. We're still experimenting with it. We're still investing and trying to do it, but it hasn't worked as well as we thought.”
Interestingly, AI has often been hailed as a champion for predicitive aircraft maintenance, with third-parties offering said services.
Kirby highlighted delays and baggage recovery as key areas that AI can potentially excel in.
For a more detailed look at past use cases for AI as well as the rise of generative AI in the sector, read our feature in our latest issue, Airline Economics Issue 85. Register for access here.