From turnaround management to predictive maintenance, AI is already moving from experimentation to daily operations at Rome Fiumicino, but human judgement remains essential for safety, accountability and resilience. Ivan Bassato, Chief Aviation Officer of Aeroporti di Roma and co-chair of the International Airport Summit roundtable on operational resilience and capacity, explains.
Beyond digital transformation
Airport infrastructure has traditionally been defined by physical assets such as runways, taxiways, aprons and terminal facilities. Today, however, a further operational layer is becoming increasingly relevant: a digital infrastructure based on data, algorithms and rapid decision-support capabilities. Artificial intelligence (AI) is a key enabler of this evolution. At Aeroporti di Roma, and particularly at Rome Fiumicino, AI is already supporting more efficient use of constrained capacity, improving disruption management and enhancing operational continuity. These capabilities are particularly relevant in aircraft turnaround operations, where several interdependent activities must be completed safely, reliably and within very limited time windows, often targeting a turnaround time of 45 minutes or less. AI-based systems are therefore increasingly used to support resource allocation, delay prediction and situational awareness. Nevertheless, these systems do not remove the need for human judgement. In abnormal or rapidly evolving circumstances, experienced operational decision makers remain essential to interpret context, balance competing priorities and maintain safety and service continuity.
The following sections examine how AI is reshaping airport operations, the regulatory principles that should govern its use, and why human oversight remains a fundamental condition for safe and accountable decision-making.
The rise of AI in airport operations
Airports generate large volumes of operational data across aircraft movements, passenger flows, weather conditions, baggage handling, security processes and ground handling activities. AI systems can process these data streams at a scale and speed that exceed manual analysis, identifying patterns, generating predictions and supporting more informed operational decisions. In this context, AI can improve the accuracy of arrival and departure forecasts, optimise gate and stand allocation, and enhance the allocation of ground resources.
At Rome Fiumicino Airport, these capabilities are already part of operational practice: Assaia ApronAI uses computer vision and existing CCTV infrastructure to monitor key turnaround events in real time, while Aeroficial Intelligence provides an airside digital twin that makes aircraft movements visible and predictable across the airfield. AI can also increase the reliability of physical airport assets. For example, sensors installed on baggage systems, power equipment or other critical infrastructure can monitor parameters such as vibration, temperature and utilisation levels.
When these data are combined with AI analytics, maintenance can shift from a reactive to a predictive model, reducing downtime and lowering the probability of failures with wider operational consequences. In turnaround management, AI supports co-ordination among multiple stakeholders by identifying potential conflicts and proposing mitigation actions before delays become unavoidable.
This moves airport operations from a reactive posture towards a more anticipatory and data-driven model. Supporting this shift is ADR’s Open Innovation model and its 650sqm Innovation Hub at Fiumicino, where startups test solutions directly in the airport environment. Over four years, more than 1,200 startups applied, over 40 PoCs were launched and 18 projects reached full-scale implementation.
ADR Ventures and the Airports for Innovation network extend this approach beyond individual pilots, helping promising technologies scale. As a result, operational precision becomes a strategic asset: the airport of the future will not be defined only by its ability to handle more flights or passengers, but by its capability to manage complexity through advanced decision support.
AI and the regulatory perspective
Aviation is founded on stringent requirements for safety, accountability and traceability. The introduction of AI in airport operations must therefore be embedded within a robust governance framework. The European Union Aviation Safety Agency, or EASA, is developing guidance on the use of AI in aviation, with particular emphasis on safety, security, transparency and human-centred design.
This approach requires that human operators remain appropriately involved, especially when AI systems contribute to operational or safety-related decisions. Core principles include explainability, accountability and data quality. These principles are consistent with broader European regulatory developments, including the EU Artificial Intelligence Act, which reinforces the need for human oversight and transparency in high-risk AI applications. In an airport context, AI can be trusted only when its recommendations are understandable, its limitations are known, and clear procedures define when human intervention is required. Compliance with these principles is what makes AI operationally useful, responsible and safe.
The limits of algorithmic decision-making
Despite significant progress, AI systems remain dependent on the quality, completeness and representativeness of the data on which they are trained, as well as on the assumptions embedded in their models. They tend to perform effectively in stable or recurrent conditions but may become less reliable when operating in highly uncertain or unprecedented scenarios. Aviation regularly involves this type of uncertainty, including sudden weather changes, technical failures, capacity constraints or unexpected disruptions.
In such circumstances, an algorithm may produce a recommendation that is technically coherent but operationally incomplete. For example, during a disrupted turnaround, an AI system may optimise for efficiency, while human operational managers may need to consider additional factors such as safety constraints, regulatory co-ordination, passenger impact or wider network implications.
Algorithms trained on historical data may also struggle with novel situations or optimise against objectives that are no longer appropriate in the specific context. These limitations do not undermine the value of AI. Rather, they confirm the need to understand the boundaries of automated decision-making and to retain effective human oversight.
Why human-in-the-loop matters
The human-in-the-loop, or HITL, principle is central to the safe and responsible deployment of AI in aviation. It requires that human operators remain actively involved in reviewing, validating and, where necessary, overriding automated recommendations. In airport operations, this principle is particularly important because controllers, operational managers and supervisors bring contextual awareness, professional judgement and an understanding of competing priorities that cannot be fully replicated by automation.
Human oversight also provides an essential layer of accountability. When decisions affect safety, punctuality, service quality or operational continuity, organisations must be able to explain how those decisions were made and who was responsible for them. Maintaining human supervision also reduces the risk of automation bias, where users may accept system-generated recommendations without sufficient critical assessment.
The most effective operational models are therefore not those that replace human expertise with AI, but those that combine advanced analytical capabilities with experienced human decision-making. At ADR, the same human-centred approach also shapes how innovation is developed: the Innovation Cabin Crew involves colleagues from different business units in mentoring startups and defining PoC KPIs, while the Hangar Program gives employees a direct role in generating and developing new ideas.
Building the airport operations centre of the future
The next generation Airport Operations Centre, or APOC, will increasingly rely on AI-enabled decision-support capabilities. However, technology alone will not be sufficient to deliver this transformation. The real value will depend on the ability to redesign operating models, decision rights and coordination mechanisms across the airport ecosystem. This requires investment not only in platforms, data architectures and algorithms, but also in the skills and governance structures needed to use them effectively. Airport operators such as Aeroporti di Roma will need to strengthen their teams’ ability to interpret data, challenge AI-generated recommendations and co-ordinate decisions with airlines, ground handlers, regulators and other stakeholders.
New capabilities will be required, including AI literacy, data-driven operational management and the ability to work effectively in dynamic and uncertain environments. Governance arrangements, escalation procedures and training programmes must ensure that AI is used in a transparent, controlled and safety-oriented manner.
Technology providers, in turn, must develop systems that are explainable, reliable and credible in real operational conditions, so that operators can trust the recommendations and act on them appropriately. Close co-operation among regulators, airports, airlines and technology companies will be essential to ensure that innovation strengthens safety, resilience and operational performance.
Operational precision is rapidly becoming a new form of airport infrastructure. At Aeroporti di Roma, this principle is already shaping the evolution of Rome Fiumicino, where investment in physical capacity goes hand in hand with digital and AI-enabled solutions designed to make operations increasingly predictable, coordinated and resilient.
AI offers significant opportunities to improve turnaround performance, optimise resource utilisation and strengthen resilience. At the same time, aviation remains a human-centred industry, grounded in expertise, judgement and accountability. As AI becomes more deeply embedded in airport operations, its purpose should be to augment human decision-making rather than replace it.
When AI performs effectively, it can improve speed, consistency and situational awareness. When AI reaches the limits of its models, human judgement remains the critical safeguard for safety, continuity and operational coherence. The airports that will succeed are those capable of combining technological innovation with disciplined governance and meaningful human oversight. In aviation, true operational excellence will depend on both intelligent systems and the professional judgement of the people who guide them.
Join me in Rome on 10-12 November 2026 for International Airport Summit where I will be co-chairing two VIP roundtable discussions; one on ‘The 45-minute turn: is operational precision the new infrastructure?’ and the other on ‘When the algorithm gets it wrong: human judgment in an ai-optimised operation’. The roundtable discussions are closed-door, invite only discussions for senior airport leaders.
We will have the opportunity to deep-dive these two topics, allowing us to speak candidly under Chatham House rules and to network with each other around our common interests and challenges.
No comments yet