Artificial intelligence is no longer a future concept in aviation. It is already embedded across a growing range of activities, from aircraft systems and air traffic management to maintenance, safety analytics and passenger processing. As the industry accelerates its digital transformation, a new question is emerging for regulators and policymakers: who remains accountable when AI helps shape operational and regulatory decisions? This issue was among the key themes discussed during the 4th ICAO Civil Aviation Legal Advisers Forum (CALAF/4), held in Nassau, Bahamas, in May 2026.
AI is already part of aviation
The debate is no longer about whether artificial intelligence will enter aviation. According to discussions presented at CALAF/4, AI is already being deployed in aircraft systems, air traffic management environments, maintenance processes, safety data analysis and passenger-facing operations.
At the same time, aviation stakeholders are exploring increasingly sophisticated applications. Future air traffic management concepts envision highly connected ecosystems supported by predictive intelligence, dynamic airspace management, automated surveillance and real-time data sharing between aircraft, infrastructure and service providers.
For airports, airlines and air navigation service providers, these technologies promise greater efficiency, improved capacity management and enhanced operational resilience. However, they also raise new legal and governance challenges.
When assistance becomes decision-making
One of the central issues raised during the forum concerns the role of AI in certification, licensing and regulatory processes.
A scenario presented during CALAF/4 illustrates the challenge: an AI system analyses data, classifies risks and generates recommendations, which are subsequently reviewed and signed by a human authority before a certificate or licence is issued. Yet if the recommendation generated by the algorithm effectively determines the outcome, what exactly is the role of the human decision-maker?
The question has important international implications. Under the Chicago Convention, certificates and licences issued by one State are recognised by others on the basis of trust in the issuing authority. If AI increasingly influences regulatory decisions, States must remain capable of explaining, validating and ultimately standing behind those outcomes.
Human oversight versus human presence
Perhaps the most significant concept discussed during the forum is the distinction between human presence and meaningful oversight.
A regulator who merely reviews an AI-generated recommendation before signing may technically remain “in the loop.” Yet forum participants argued that genuine oversight requires more than a procedural approval step. It requires the ability to independently assess information, exercise judgement and defend the final decision.
As aviation systems become increasingly automated, regulators may need to determine where meaningful human control begins and ends. This debate is already emerging beyond aviation.
Examples presented during CALAF/4 highlighted how governments and regulators around the world are approaching the issue. The European Union’s AI Act requires human oversight for high-risk AI applications, while U.S. aviation regulators continue to emphasise that technology can support safety-critical decisions but should not replace human judgement.
A challenge for Latin America and the Caribbean
The issue is particularly relevant for Latin America and the Caribbean, where airports, airlines and aviation authorities are investing in digitalisation initiatives, smart airport technologies and modernised air traffic management systems.
As these technologies become more deeply integrated into operational and regulatory environments, governance frameworks will need to evolve alongside them. The challenge is not simply technical. It concerns accountability, transparency and trust across an international aviation system built on shared standards and mutual recognition.
Looking ahead
The discussions at CALAF/4 suggest that the next phase of aviation digitalisation will be shaped as much by legal and regulatory considerations as by technological innovation. Artificial intelligence is already becoming part of the industry’s decision-making architecture. The question facing regulators is no longer whether AI should be used, but how responsibility should be allocated when humans and machines make decisions together.
For aviation authorities, airports and industry stakeholders, ensuring that human oversight remains meaningful may become one of the defining governance challenges of the decade.
