A delayed aircraft arrives while the next departure is already boarding. A baggage conveyor needs attention. Queues are growing at security. For the people managing an airport, the challenge is deciding which problem needs action first, with limited time and an incomplete picture of what happens next.
Artificial intelligence promises to improve that picture. By identifying patterns, forecasting demand and flagging emerging problems, it can help teams intervene earlier and more effectively. But any prediction is only useful when someone understands it, trusts it appropriately and knows how to make the best use of it.
As airports explore new and emerging AI solutions, that relationship between technology and people deserves greater attention, particularly given the wider political and economic complexities of the facilities themselves. Airports aren’t just regular businesses; they represent national borders, and (inevitably) centres of interconnected national interests. That makes the practical questions of AI’s impact on airports even more crucial. Beyond the surface-level question of what can an AI system automate or otherwise improve: how should employees learn to use it, which responsibilities will change, and how can airport leaders prepare their operations and infrastructure for the years ahead?
Where AI is already making a difference
Some of the most useful airport applications address the perennial operational frustrations that reach back to the dawn of commercial flying.
At Amsterdam Airport Schiphol, Deep Turnaround uses cameras and AI to translate activity around aircraft stands into operational data. It provides insight into more than 30 turnaround processes, including cleaning, catering and refuelling, helping teams identify delays and predict when an aircraft will be ready. Schiphol developed the solution through discussions with airlines, ground handlers, gate planners and air traffic controllers.
The significance is practical, as earlier information gives teams more time to respond. A gate planner can assess the implications of a late turnaround, while a ground handler can investigate the activity holding it up. Their experience remains essential to deciding what action is feasible.
Closer to home, Dubai Airports offers another prime example. Its AI and machine-learning-powered material requirements planning tool forecasts spare parts consumption, helping engineering teams obtain the parts they need. In its published account of the initiative, Dubai Airports reported a 30% improvement in forecast accuracy and a 12% reduction in excess stock.
Saudi Arabia has also taken steps towards AI-assisted security screening. In a report published on 30 September 2025, the Saudi Press Agency described AlUla’s air traffic being managed remotely through a virtual control tower using high-definition cameras, advanced sensors and AI systems for data analysis and airspace monitoring.
These examples involve different forms of AI. Computer vision interprets images; predictive models estimate future events; generative AI produces text and other content that can help establish the vision for better operations and even whole systems. Airport leaders need to understand those distinctions because the training, validation and operational risks differ accordingly.
Start with an operational problem
For airports considering their next investment, the first task is to define a problem precisely enough to measure tangible improvement.
For example, “improve passenger flow” is a broad ambition. Predicting when queues will exceed an agreed threshold, with enough warning to open another lane, is a clearer starting point. The same principle applies to maintenance, stand allocation or administrative work. When airports identify the decision, the person responsible and the information they need, a suitable AI solution will be easier to establish.
A useful pilot should therefore determine a baseline and test performance under realistic conditions. Does the forecast remain reliable during disruption? How many unnecessary alerts does it generate? Can staff act on the information quickly enough to benefit?
Data preparation is an essential part of that work. Inconsistent timestamps, incomplete maintenance records or incompatible partner systems can undermine an otherwise promising model. Airports should also agree who can access the data, how it will be protected and how the service will operate if the AI becomes unavailable – a wise precaution given the fact that AI is such a nascent industry that is attracting growing global geopolitical interest.
Involving frontline employees early helps expose these issues. They know which records are reliable, where procedures differ from practice and why an apparently sensible recommendation might fail during a busy shift. As always, AI shouldn’t be viewed as a simple replacement for human intelligence and judgement. Rather, it should reinforce them.
Train employees for the decisions they will make
AI training should reflect the work people actually do. A general awareness session can introduce the technology, but it cannot comprehensively prepare a security officer, maintenance planner and commercial analyst for their different responsibilities.
Changi Airport Group provides a useful example of building capability alongside adoption. Its 2024/25 annual report describes workshops developed with its in-house Centres of Excellence on data analytics, AI and robotic process automation. It also describes designated innovation days that give employees opportunities to learn and experiment.
For other airports, a practical programme could begin with shared foundations: what approved tools can do, where they can be wrong, which information employees may enter and when human review is required. Training should then move into specific roles.
An operations controller needs to interpret a prediction and understand when to override it. A maintenance planner needs to assess an alert against equipment history and engineering judgement. An employee using generative AI to draft passenger information needs to check facts, language and instructions before publication.
This is a reminder that practice matters. Teams should rehearse situations in which a prediction is wrong, an alert arrives late or a system becomes unavailable. Knowing how to challenge a recommendation and return to established procedures is as important as knowing how to use the interface.
Training also needs protected time, support during shifts and refreshers when tools change. Attendance alone is a weak measure of readiness. Demonstrating that employees can use a system correctly, recognise its limitations and escalate concerns provides a stronger basis for deployment.
How will airport jobs change?
Employment is likely to be the most personal concern for staff, and leaders should address it directly.
The International Labour Organization’s 2025 research found that one in four workers globally is in an occupation with some exposure to generative AI. Its assessment was that transformation of jobs is more likely than complete replacement, because many tasks still require human input. This is a global assessment of generative AI, rather than a forecast for airport employment.
For airports, examining tasks offers a more useful starting point than predicting the disappearance of entire roles. Routine reporting, document searches and initial drafting may become faster. Operations staff may spend less time gathering status updates and more time coordinating responses. Engineers may receive better information about assets while retaining responsibility for diagnosis and safe intervention.
New responsibilities could include checking data quality, monitoring model performance and translating operational needs for technology teams. Airports should consider how existing employees can develop into these responsibilities, alongside any specialist recruitment.
However, redeployment is not automatic. Some tasks may shrink substantially, and the effects will vary between operators, contractors and departments. Workforce planning should identify those changes early, establish retraining routes and discuss them with employees and their representatives. Faster work should also prompt questions about workload, service quality and staffing, rather than an assumption that every efficiency gain justifies a headcount reduction.Â
Bringing practical questions to GAF 2026
For airport professionals, the next step is to explore the technology with clear operational questions. What has a solution achieved in a working airport? What data does it require? How does it fit existing systems? What training and ongoing support will employees receive?
Global Airports Forum 2026, taking place in Riyadh between 30 November and 1 December, offers an opportunity to explore AI-enabled solutions and discuss these questions with technology providers and industry peers.
Those conversations can help airport teams assess where AI is heading and what it will take to make it useful in their own organisations. The strongest investments will connect better information with people who are prepared to act on it.
