Has UK SORA Made Drone Safety Better— or Just More Expensive?

The UK drone industry does not need less safety. It needs better safety regulation.

That distinction matters. Most serious operators accept that complex drone operations need proper risk assessment, competent pilots, technical controls, emergency procedures and clear operating limits. Nobody sensible is arguing for a free-for-all. The problem is that the Civil Aviation Authority’s move from the old Operating Safety Case system to UK SORA risks replacing practical safety management with a process that is more expensive, more document-heavy and increasingly detached from the reality of day-to-day drone operations.

The old OSC system was not perfect. It could be inconsistent, and it relied heavily on the quality of the applicant’s own safety case. But it had one major advantage: it was understandable. CAP 722A was built around three main volumes — Volume 1, the Operations Manual; Volume 2, the UAS Description; and Volume 3, the Safety Risk Assessment. The CAA described the OSC as the operational risk assessment used to support an application for Specific Category authorisation, with the aim of showing that hazards had been identified and reduced to a tolerable and ALARP level. (Civil Aviation Authority)

SORA, by contrast, has turned the process into a far more fragmented compliance exercise. EASA describes SORA as a 10-step risk assessment process that considers ground risk, air risk, SAIL level, containment and 24 Operational Safety Objectives, or OSOs. The higher the SAIL, the more demanding the required level of robustness and evidence becomes. (EASA)

On paper, that sounds systematic. In practice, it means operators now have to manage a growing evidence web: an operations manual, technical files, OSO evidence, containment evidence, population assumptions, aircraft data, training evidence, maintenance evidence, emergency procedures, ground-risk mitigations and, in many cases, manufacturer or RAE(F) evidence. The CAA itself confirms that operators still need an operations manual under UK SORA, but may also need to provide separate compliance evidence such as technical information. (Civil Aviation Authority)

That is the core problem. The industry has not moved from “three OSC volumes” to “a simpler digital process”. It has moved from a consolidated safety case into a document-tracking exercise where each safety claim needs its own supporting evidence.

The flawed foundation: population data

One of the most questionable parts of SORA is the way it relies on population density to determine ground risk. The CAA says operators need the maximum population density in the operational volume and ground risk buffer, and UK SORA guidance says population density should be determined using maps with an appropriate grid size. It lists ONS Census data, ESA Copernicus data, applicant survey data and other verified resources as possible sources. (Civil Aviation Authority)

That may appear objective, but it is often not operationally realistic.

Population density is not fixed. It changes by time of day, season, weather, school holidays, events, road closures, tourism, local work patterns and even whether it is sunny or raining. A park, beach, seafront, town square or event site can move from almost empty to crowded in hours. A static population layer may be useful for broad planning, but it can be a poor proxy for the actual number of people exposed during a real flight.

The ONS itself states that the census takes place every 10 years, and that small-area population estimates are subject to statistical uncertainty, especially in small areas, high-churn areas and towards the end of the inter-census period. (Office for National Statistics) The ONS also revised mid-2022 and mid-2023 population estimates after incorporating improved migration and administrative data, showing that even official population estimates change after publication. (Office for National Statistics)

This does not mean ONS data is poor. It means it was not designed to be treated as a precise, live safety input for every drone operation. SORA can therefore create a false sense of accuracy: a risk score looks scientific because it uses numbers, but the numbers may not represent the actual operating environment.

Weather makes this even more obvious. Weather affects aircraft performance, visibility, pilot workload, emergency response and where people are likely to be on the ground. EASA’s own UAS rules recognise that operating procedures should consider environmental and weather conditions and methods of obtaining forecasts. (EASA) The old CAP 722A OSC template also required operators to describe how weather forecasts and conditions would be obtained so that UAS operating limits were not exceeded. (Civil Aviation Authority)

So why does SORA lean so heavily on a population model that can be static, stale or only loosely representative of the real site? A drone operation is not conducted inside a spreadsheet. It is conducted in weather, in public spaces, around changing people flows and with real operational constraints.

Complexity has created a consultant-dependent process

SORA was intended to standardise risk assessment. In reality, it has become so complex that many operators cannot realistically complete it without specialist support.

That is bad for innovation. A safety regime should reward competent operators and well-designed aircraft. It should not primarily reward those who can afford consultants, document writers and repeated regulatory submissions.

The CAA itself recognises that the transition to UK SORA is “a big change for operators” and offers support calls and webinars. (Civil Aviation Authority) That is a telling admission. If a process is so complex that the regulator needs to run repeated guidance sessions just so industry can understand how to apply, then the system has already drifted away from proportional regulation.

The risk is that operators spend more time proving that they understand SORA than proving they can actually operate safely.

The cost comparison is stark

The old OSC route was not cheap, but the cost base was at least relatively predictable. In the 2024/25 CAA charging scheme, a Specific Category Operational Authorisation with an Operating Safety Case or risk assessment was listed at £2,185 for a standard initial application, with a £625 renewal. A reduced “Case 2” application was £1,249, also with a £625 renewal. (Civil Aviation Authority)

Under the current UK SORA charging structure, the numbers are much higher. For 2026/27, the CAA lists initial SORA charges of £2,422 for SAIL 1, £3,806 for SAIL 2, £10,380 for SAIL 3, £13,840 for SAIL 4, and £17,300 for SAIL 5 or 6. The authorisation lasts 12 months, and additional charges may be applied at £346 per hour where the fixed fee does not cover the assessment time. (Civil Aviation Authority)

That means a SAIL 2 application is around 74% higher than the old standard OSC initial fee, before considering consultant time, technical evidence, document maintenance or any RAE(F) involvement. SAIL 3 is nearly five times the old standard OSC initial fee. Compared with the old £625 renewal, the difference is even more severe.

For a large organisation, this may be manageable. For a small operator, manufacturer or start-up, it is a barrier to entry.

SAIL-marked aircraft are not a simple answer

The theory is that SAIL-marked aircraft will make compliance easier. In time, that might be true. But today it adds another layer of difficulty.

The CAA says RAE(F)s assess whether the technical features of a UAS meet agreed UK SORA requirements, including design, construction and flying characteristics. It also says designers may need to use an RAE(F) to obtain a SAIL Mark Certificate, and that SAIL Mark Certificates map to SAIL levels 1 to 6. (Civil Aviation Authority)

That creates a new dependency chain. Operators need compliant aircraft. Manufacturers need SAIL marking. RAE(F)s need approval. Evidence has to be generated, assessed, maintained and paid for. The CAA’s own 2026 scheme lists an RAE(F) initial application charge of £9,800, with excess charges at £346 per hour. (Civil Aviation Authority)

This is not a lightweight ecosystem. It is a regulatory supply chain.

For manufacturers, especially smaller UK companies, the burden is obvious. If an aircraft is already built from standard components that comply with relevant radio, electrical and product standards, a further flightworthiness assessment may still be required for SORA purposes. That may be justifiable for genuinely complex or high-risk operations, but it becomes questionable when the assessment largely duplicates evidence that already exists in component certifications, conformity documentation, design files and operational testing.

The danger is that the UK creates a process where new aircraft are not blocked because they are unsafe, but because the cost and time needed to prove compliance becomes too heavy.

Paperwork is not the same as safety

The most important point is this: more documents do not automatically mean safer operations.

A well-run OSC could show the aircraft, the operation, the crew, the procedures and the risk controls in a coherent way. SORA risks breaking that same safety argument into multiple compliance fragments. Each fragment may be logical on its own, but the total burden becomes disproportionate.

Operators now have to maintain an operations manual, risk methodology, OSO evidence, technical declarations, ground-risk assumptions, population data, containment evidence, training records and potentially SAIL-marked aircraft evidence. That might be suitable for high-end BVLOS operations or large aircraft operating in complex airspace. It is much harder to justify as the default direction of travel for all serious Specific Category operators.

The result is predictable: operators spend more time maintaining paperwork than flying, testing, training and improving real-world safety.

What the CAA should change

The answer is not to scrap safety cases. The answer is to make the system proportionate..

Since SORA was designed, drones have gotten safer and safer to operate and more reliable, and this happened all before SORA regulation came along. After all over flight of people is not unsafe, so maybe its time to adopt the ‘do not stop and hover directly over uninvolved 3^rd parties’

For aircraft, the CAA should avoid turning SAIL marking into a quasi-certification regime for every manufacturer. If standard components already meet recognised requirements, and the aircraft has documented design controls, maintenance procedures, flight logs, containment features and operational limits, the assessment should focus on the actual additional risk — not simply re-checking paperwork for its own sake.

The safety record should matter

One point often missing from the SORA debate is the actual safety record of civilian drone operations.

CAP 722 has existed since 2002, and multirotor drones have since become a normal part of UK aviation, public safety, surveying, media, inspection and event operations. Yet there is no clear evidence of a fatality in the UK/EU/USA caused by a commercially operated civilian multirotor drone working under the CAA framework. That does not mean the risk is zero, and it does not mean regulation is unnecessary. But it does mean the regulatory response should be proportionate to the demonstrated risk, not driven by theoretical worst-case modelling alone.

The CAA’s own 2024 safety review shows a rapidly growing drone sector, with around 220,000 active registered drone flyers and operators, more than 2,500 active Specific Category Operational Authorisations, and 55 RPAS accidents or serious incidents reported during 2024. That is not a record of an industry causing widespread public harm. It is a record of an industry that has grown while still showing a very low level of severe consequence events.

This matters because SORA often treats low-probability ground-risk scenarios as though they are the natural starting point for regulation. The process begins with modelled risk, population assumptions and consequence tables, then demands layers of documentation to reduce that modelled risk. But if more than two decades of civilian drone operations have not produced the fatality profile implied by the paperwork burden, then the system should be asking a different question: are we regulating against the real operational risk, or against an abstract version of the risk?

A mature regulator should use both leading and lagging indicators. It should not wait for accidents before acting, but neither should it ignore the fact that the existing industry has operated for years without the catastrophic public safety outcomes often used to justify increasingly complex regulation.

Conclusion

The irony is that SORA has arrived after more than two decades of civilian UAS regulation under CAP 722, during which the UK has not seen the kind of fatal multirotor accident record that would justify treating paperwork complexity as a substitute for practical, proportionate safety management.

If the process relies on static population data that may not reflect the real operating environment, requires operators to maintain a sprawling evidence library for every OSO, pushes manufacturers towards difficult and expensive SAIL marking, and increases costs from thousands to many thousands of pounds, then the industry is right to question whether this is making operations safer or simply making compliance more expensive.

The UK drone sector needs regulation that supports safe innovation. At the moment, SORA risks doing the opposite: creating a paperwork-heavy, consultant-dependent system where the burden of proving safety starts to outweigh the practical work of being safe.

By

Not a UAS Consultant