Is aviation law set for automatic take-off?

Recent breakthroughs in aviation have seen the development of innovative autonomous and highly automated systems and vehicles, such as drones, advanced air mobility vehicles and vertical take-off and landing aircraft (VTOLs) which can provide short journeys for a small number of people.

While acknowledging that increased automation has the potential to deliver substantial benefits to the entire aviation system, UK industry, and the public, to realise these benefits the UK’s legislative and regulatory framework must be sufficiently agile to facilitate innovation, whilst robust enough to maintain the high safety standards that aviation enjoys.

The pace of innovation can be rapid, and the technological functions that are developed are often unforeseeable. This gives rise to a classic challenge for law makers: how to anticipate, and respond to, advances in technology.

Law Commission project

The legislative response is ultimately a question of policy. However, there are instances where the Government seeks neutral, technical, evidence-based input from the UK’s Law Commissions, that in turn seek views from industry and the legal profession through consultations. The Law Commission of England and Wales is leading a project in relation to autonomous flight. The initial consultation in relation to drones and VTOLs closed in June 2024 and a second consultation paper is expected in early 2025, followed by a final report to Parliament in 2025.

The review builds on the framework for automated vehicles within the Automated Vehicle Act 2024 that received Royal Assent in May 2024 (but is not yet in force). There are certain overlaps.

What is an unmanned aircraft?

For autonomous flight there are unique challenges around classification that do not arise in relation to automated cars. While it is tempting to divide ‘unmanned aircraft’ into ‘autonomous aircraft’ and ‘remotely piloted aircraft’, this is a difficult distinction to draw as aircraft are already not entirely one or the other.

 Autonomy is a sliding scale, with the involvement of pilots on the ground varying in nature and degree. Flight, even in aircraft with human pilots, is already highly automated, and in relation to unmanned aircraft, the involvement of a human pilot on the ground can take many forms. There will be instances where the human pilot takes control, has oversight, or has no involvement at all. When things go wrong, a pilot may intervene and indeed may have to be capable of intervening if new aircraft are to gain certification.

This brings to mind my jurisprudence lectures at The University of Edinburgh in the early 1990s and questions of legislative interpretation. HLA Hart gave an example of a rule: “No vehicles in the park”.  He then asked, is a skateboard a vehicle?  What of a World War II tank installed as a war memorial? These “hard cases” seem straightforward by comparison!

Questions of liability

The challenges of defining responsibilities and allocating liability are obvious. They also have a certain novelty in relation to unmanned aircraft. Unlike the traditional aviation industry where manufacturers and operators are separate, innovative vehicles will likely be made and operated by the same companies.

What of machine learning?

The rules of flight have evolved through human experience and rely on the fact that the pilot (on board) has ultimate responsibility for safety. In emergency scenarios pilots must be (and are) able to override the rules of flight in situations that render such departure absolutely necessary in the interests of safety.

In the era of machine learning and artificial intelligence, should autonomous aircraft be allowed to do the same and, if so, what are the implications? What if things go wrong – how should the law allocate civil and criminal liability?

For software to override the rules of flight in pre-conceived situations is one thing, but should machines that are capable of learning from experience be allowed to exercise autonomy in scenarios that have not be preconceived? If so, where should the law attribute responsibility when things go wrong?

Putting questions of capability and culpability to one side, a deeper question arises: even if aircraft develop the capability to operate with autonomy, and to learn from experience, will the public accept this?

Drones

Drones have applications and potential applications in many realms - emergency services, telecoms, surveillance, agriculture, making deliveries, transport, and infrastructure to name a few.  But what exactly is a drone? There is no universally accepted legal definition, but the Commission’s working assumption is that drones are smaller, unmanned, and unoccupied aircraft that can either be remotely piloted or autonomous.

The current legal position is that drones can only be flown within the visible line of sight (VLOS) and that any flight that is beyond the visible line of sight (BVLOS) requires authorisation from the Civil Aviation Authority. As technology develops, legislators must build in some flexibility so that undue pressures do not build up on regulators.

Air traffic management and air management systems

There are also implications arising from the fact that drones and VTOLs will require information and traffic-management services that traditional aircraft simply do not require, given the marked differences in weather and topography at lower levels and in urban areas.

Hijacking

And how do we update our laws around hijacking? The current law states that ‘a person on board an aircraft in flight who, unlawfully by the use of force or by threats of any kind, seizes the aircraft or exercises control of it commits the offence of hijacking.’  (Aviation Security Act 1982, section 1(1).)

The International Civil Aviation Organisation has agreed a new protocol to include ‘by digital means’.  The UK has signed this protocol but not ratified it, so it is not yet part of UK domestic law. The reference to the aircraft being ‘in flight’ also merits attention – what if control is seized in advance of that?

Next steps

The Law Commission seems likely to recommend that the Government sets up an industry forum to allow collaboration on how the rules of flight, air traffic management and air management services can adapt to new technologies, just as it has done in relation to automated vehicles.  

This would be a sensible move. However, in terms of the certification of new aircraft, it seems highly ambitious to think that the law will facilitate innovation.  Certification of new vehicles and aircraft takes years and is necessarily reactive. Similarly, designing schemes of liability to govern new technologies operating in new ways is difficult to do in advance.  

Written by Ralph Riddiough, director at Holmes Mackillop solicitors