By Alex Stovold & Lorraine Wilson
|Published 01 September 2026
As semi-autonomous and autonomous aircraft move steadily towards operational use, questions arise as to whether existing frameworks of international aviation law are capable of accommodating new forms of air transport. In particular, does the Montreal Convention 1999 ("the Montreal Convention")[1] require amendment to address passenger, baggage and cargo claims that may arise from the future introduction of autonomous aircraft into international air carriage?
The issue of legal liability arising from autonomous air transport was raised at the 42nd Assembly of the International Civil Aviation Organization ("ICAO") in 2025, including through a working paper submitted by Kazakhstan calling for consideration of a harmonised international framework governing liability and compensation in autonomous aviation.[2] The proposal was made for a dedicated legal study group to explore a harmonised international instrument, whether a new convention or a supplementary protocol, to clearly define liability and compensation mechanisms in the context of what the working paper described as the "new frontier" of autonomous aviation. The working paper concluded that "the future of civil aviation will undoubtedly involve increasing levels of autonomy and machine intelligence. The legal system must evolve in parallel to maintain safety, accountability, and justice."
Although the proposal highlighted the need for further study, it was not translated into a new work item within ICAO's legal programme and there is presently no indication that ICAO intends to pursue the matter before the next ordinary Assembly in 2028. The Extraordinary Session scheduled for 2026 is confined to governance matters and is not expected to address autonomous aviation liability.
Meanwhile, the 40th Session of the ICAO Legal Committee, due to take place in Montreal in October 2026, will consider reports on a number of items within its General Work Programme.[3] That agenda includes consideration of international legal aspects of unmanned (pilotless) aircraft operations and their integration into civil aviation. Although the scope of that work is broad enough to encompass increasingly autonomous forms of aircraft operation, it is not presently directed in express terms towards reform of the Montreal Convention or the development of a dedicated liability regime for international autonomous air transport. The position is therefore clear: autonomous aircraft operations are already recognised by ICAO as a significant legal and regulatory issue. At this stage, the discussion is centred on how such aircraft can be integrated into the civil aviation system rather than on whether the Convention liability regime requires reform and what any reform might look like.
Historically, ICAO has favoured gradual development rather than fundamental reform. Matters appearing on the Legal Committee's agenda will typically have been the subject of extensive analysis, working group activity, and international consultation before any formal proposals emerge.
The Montreal Convention itself provides a useful illustration. It was not conceived and adopted within a single ICAO triennium. Rather, it emerged from several years of Legal Committee work and international consultation before being adopted at the International Conference on Air Law in Montreal in 1999. By the time States met in Montreal in 1999 to adopt a new convention, international air carrier liability had already been shaped by a succession of instruments and amendments, including the Warsaw Convention 1929, the Hague Protocol 1955, the Guadalajara Convention 1961[4] and a series of subsequent additional Protocols.[5] The Convention was therefore not a revolution so much as the culmination of a process of legal evolution aimed at replacing an increasingly fragmented framework with a single, modern legal liability regime.
Autonomous aviation is likely to follow a similar path. The fact that pilotless aircraft now feature within ICAO's legal work programme should not be mistaken for a sign that legal liability treaty reform is imminent. Rather, it reflects ICAO's longstanding practice of identifying legal and regulatory challenges well before they arise in practice.
The immediate priority continues to be understanding the issues, developing an appropriate regulatory framework and gaining operational experience before considering whether more fundamental legal reform is required. That has long been ICAO's approach: operational experience first, legal reform later.
Recognising autonomous flight operations as a juridical and regulatory challenge is one thing. Deciding that the Montreal Convention requires amendment is quite another. Nevertheless, autonomous commercial flight is no longer confined to research programmes and demonstration projects. Pilotless passenger-carrying aircraft have already received approval for limited commercial operations in some jurisdictions[6], while cargo operations are widely regarded as the most likely pathway towards broader deployment of highly automated air transport. Although widespread autonomous commercial operations are unlikely in the near term, autonomous aviation can no longer be dismissed as a purely theoretical prospect.
Yet for all the attention being given to autonomous aircraft operations, reform of the Montreal Convention is notably absent from ICAO's agenda. ICAO's current priorities remain centred on promoting universal ratification of the Convention, preserving its role as the global liability framework governing international carriage by air, and administering periodic adjustments to liability limits through the Article 24 review mechanism. The most recent increase in liability limits, which took effect on 28 December 2024, was implemented through that mechanism and did not require amendment of the treaty itself.
A further reason for ICAO's measured approach may be that autonomous aviation is likely to develop domestically before it develops internationally. Initial approvals for passenger and cargo operations are likely to be granted by individual States for operations conducted wholly within their own territories, whereas the Montreal Convention is concerned with international carriage by air. The legal pressures that typically drive treaty reform may therefore not emerge until autonomous operations become established on international routes.
The absence of reform proposals may also reflect the fact that the Montreal Convention is more technology-neutral than is sometimes appreciated. The Convention regulates the liability of carriers rather than the technology through which carriage is performed. Liability attaches to the entity undertaking the international carriage by air, whether as a contracting carrier or actual carrier. That carrier-focused approach may explain why ICAO has not yet concluded that autonomous flight operations require immediate amendment of the Convention.
Many of the Convention's core liability provisions therefore appear capable of continuing to operate without modification, at least for the time being. The more difficult questions may lie elsewhere. If an autonomous system contributes to an accident, to what extent should responsibility rest with the carrier, the aircraft manufacturer, the software developer, or the provider of an AI system? How do you regulate the operation of and liability for a system where an autonomous aircraft is flown through a combination of onboard, cutting-edge computing, computer vision, and machine learning models? These are undoubtedly difficult questions, but they are not necessarily Montreal Convention questions. Initial litigation arising from autonomous operations may prove to be less concerned with the passenger-carrier relationship and more focused on questions of technology, product performance, software design, and the allocation of responsibility among those involved in the design, manufacture, and operation of aircraft.
That is not to say that amendment of the Montreal Convention will never become necessary. If highly autonomous commercial air transport operations become established and courts or regulators in different jurisdictions begin reaching inconsistent conclusions regarding accidents involving autonomous systems, pressure may build for an internationally harmonised solution.
Concerns of that kind are already beginning to emerge. In its final report on aviation autonomy, published in May 2026, the Law Commission of England and Wales identified the potential need for future international engagement concerning the application of existing aviation liability frameworks to autonomous operations.[7] While this falls well short of a proposal for immediate treaty reform, it illustrates how questions regarding the adequacy of existing liability regimes may become more prominent as autonomous commercial aviation develops.
ICAO is actively examining the legal implications of pilotless and increasingly autonomous aircraft through its Legal Committee work programme and related RPAS initiatives. However, there is presently no indication that it intends to amend the Montreal Convention 1999 or develop a distinct liability framework for autonomous commercial air transport.
That reflects a degree of confidence in the existing framework. The Convention was drafted around the relationship between passenger and carrier rather than the means by which an aircraft is operated, and many of the challenges posed by autonomy may therefore be accommodated within its existing structure. For the time being, ICAO's priority remains the safe integration of autonomous technologies rather than liability reform. Its longstanding preference for responding to operational developments rather than anticipated technological change suggests that any amendment of the Montreal Convention is likely to follow, rather than precede, the widespread use of autonomous aircraft in international carriage.
[1] The Convention for the Unification of Certain Rules for International Carriage by Air, done at Montreal on 28 May 1999 (commonly known as the Montreal Convention 1999 or MC99).
[2] 42nd Session of the ICAO Assembly (Montréal, 23 September-3 October 2025), Working Paper submitted by Kazakhstan, Harmonization of Liability Laws for Aviation Authorities in the Era of Autonomous Flights: Defining Liability in AI-Operated and Unmanned Commercial Air Transport (A42- WP/300, 2025).
[3] ICAO Legal Committee, 40th Session Provisional Agenda (LC/40-WP/1-1) 16 July 2026.
[4] Convention for the Unification of Certain Rules Relating to International Carriage by Air (Warsaw, 12 October 1929); Protocol to Amend the Convention for the Unification of Certain Rules Relating to International Carriage by Air (The Hague, 28 September 1955); Convention, Supplementary to the Warsaw Convention, for the Unification of Certain Rules Relating to International Carriage by Air Performed by a Person Other than the Contracting Carrier (Guadalajara, 18 September 1961).
[5] Montreal Additional Protocols Nos. 1–4 (Montreal, 25 September 1975).
[6] EHang Holdings Limited, "EHang's EH216-S eVTOL Operators Obtain Air Operator Certificates" (30 March 2025), announcing the grant by the Civil Aviation Administration of China of the first Air Operator Certificates for civil passenger-carrying pilotless aerial vehicles, authorising commercial passenger-carrying operations including sightseeing and tourism flights.
[7] Law Commission of England and Wales, Aviation Autonomy: Final Report (Law Com No 426), May 2026.
