Autonomous defence technology: Exquisite but isolated

A recent AtkinsRéalis and Royal Aeronautical Society roundtable explored how the UK can keep pace with the rate of change in autonomous technology.

For decades, a common question was: Will autonomous systems ever be used on the battlefield? That question has now been answered, with increasingly regular deployments around the world.

Today, there are new questions: just how widespread will their use become, and is the UK keeping pace?

An AtkinsRéalis and RAeS roundtable tackled those questions with experts at the Farnborough International Airshow, exploring whether the UK should move beyond a "human-in-the-loop" approach towards greater levels of autonomous decision-making.

Who takes the risk?

For David Clark, Global Market Lead for Defence at AtkinsRéalis, "hearts and minds" are the starting point. Any shift in how autonomous systems are deployed will require social licence from the British public.

"As defence spending increases, governments need the social licence to make those difficult decisions, but that trust also depends on demonstrating that the systems being funded are safe, effective and deliver the deterrence the public expects," explained Clark.

That licence is hardest to maintain when things go wrong. Autonomous systems introduce a "different dimension" to risk, said Commodore Neil Mathieson, Head of Regulation and Certification at the Military Aviation Authority (MAA).

"Who is the person taking the risk?" he asked. The answer could include everyone from mission commanders and operators to regulators and manufacturers.

Full autonomy and "human-in-the-loop" operation can pull in opposite directions. Autonomous systems "exponentially" increase the number of decisions required of human operators, noted Andrew Munday, Global Technology Director at Babcock International.

"The quality of those decisions is going to go down," he said. "At what point does it tip over from being more risky rather than less risky?"

On whether current frameworks can support safe deployment, Mathieson pointed to a practical limitation: UK geography. Britain is "simply not big enough" to test some new long-range systems, so the use of ranges in Canada and Australia is being explored.

Development moves fastest when industry and frontline commands engage regulators early, Mathieson added.

Far from being a barrier to innovation, effective regulation can help accelerate development when brought into programmes at an early stage. Mathieson argued that collaboration between industry, operators and regulators is often the fastest route to deployment, noting that regulators are typically focused on finding solutions rather than creating obstacles. "Nine times out of ten, we will find them a route to yes," he said.

"Regulation has never stopped me, with any of my experimental aircraft," observed Lucy Crowther, Chief Engineer at Defence Equipment & Support. “We have turned around systems from idea to reality within eight weeks.”

Artificial intelligence and additive manufacturing are accelerating development and production, Munday said.

“What we are seeing is the democratisation of technology, the democratisation of producing products.”

Defence start-ups, nonetheless, often try to move too fast, said Dr Richard Drake, General Manager of Anduril UK.

“You need to be safe and do the right level of work, and you need to respect the regulator,” he said. “Be respectful of the process, and from that we can change the process together."

We do not need toys

AEROSPACE Editor-in-Chief Tim Robinson, who moderated the panel, suggested the UK can tend to develop “exquisite but isolated” platforms: advanced weapons technology that lacks interoperability or scalability in wartime

Ukraine has made the cost of fragmentation concrete.

“Ukraine flooded the battlefield with 180 different types of drones in the first few weeks,” Mathieson said. “None of which spoke to each other, none of which were integrated into the operational concepts and doctrine.”

Not all autonomous systems should be treated alike, Clark added. Exquisite, attritable and low-cost capabilities each demand a different approach to support and development. Fast-cycle systems should not be slowed by requirements they do not need.

As geopolitical tensions continue to rise, the next 12 months will be decisive for autonomous deployment.

The discussion on autonomy highlighted a broader challenge facing Western defence programmes: the ability to move from successful demonstrations to operational scale.

While the UK has shown it can develop advanced autonomous technologies, panellists questioned whether industry can manufacture and deploy them in the volumes that modern conflict may demand.

Reflecting on the pace and scale of deployment seen in Ukraine, Mathieson remarked: "Where we are lagging... is I don't have a hangar full of millions of them." As noted throughout the discussion, future advantage will depend not only on technological sophistication, but also on industrial capacity.

“We do not need toys; we need capability,” added Munday, as weapon systems can quickly become redundant in the rapidly shifting battlefield.

This article was originally published in Tech UK.