AI can unlock the superpower potential of Nato – but what are the difficulties?
Artificial intelligence is making its way into societies and armed forces alike, transforming what can be achieved with swathes of data.
To unlock the "superpower" potential of alliance membership, defence and industry across the UK and partner nations must pull their data together to paint a full picture for military platforms using AI, a former commander has said.
Retired Rear Admiral Alex Burton, former commander UK Maritime Forces, told Forces News AI algorithms would need to be shared more broadly and data-sharing must occur at a volume and pace unseen up to now.
AI's application in defence is being explored within the Future Combat Air System, drone swarming and rapid decision-making recommendations in crisis scenarios.
There have also been calls for AI models to anticipate which stocks would run dry in the opening weeks of a war, or where a vehicle would need repairing further down the line.
However, sharing data and finding common ground among allied nations' AI models is likely to become as crucial as standardising ammunition.
Will Blyth is CEO of Arondite – a UK-based firm "focused on building software and AI to connect up the battlefield".
Sharing an algorithm is "as simple as sharing any other form of file", he said, but ensuring all parties are comfortable with its use, once shared, is more difficult – as is installation to make sure it soaks up sensor data correctly.
"Sometimes that involves careful and selective declassification of data," said Mr Blyth, adding that industry partners involved often want assurances their shared data won't be copied by their competitors.
This is where RAdm (Ret'd) Burton believes China and other adversaries trump the UK and its allies.
"China's industry is largely government-owned and so there is a very porous relationship between their industrial data providers, their military data providers," he said, which offered a "data advantage".
There are fewer examples of AI in defence that have been designed from the ground up to be used between allied nations.
Aukus – a trilateral defence pact between the UK, the US and Australia – has developed a "trilateral algorithm" to share intelligence, gathered by submarine hunter-launched sonobuoys, between the three nations.
Dr Simona Soare, a senior lecturer at Lancaster University, says developing a joint AI architecture for the Future Combat Air System became a "fairly long and arduous process".
"Now we're there, there's probably going to be some differences between different member states using the platform but, essentially, it's based on a shared architecture," she said.
"That is important, I think, when it comes to the actual interoperable use of these."
However, data-sharing also offers new challenges within alliances.
Since Turkey's purchase of the Russian S-400 air defence platform, Nato has removed the country from the F-35 jet programme over fears sensitive data on the jets could be sent straight to Moscow through the missile platform.
"We're just scratching the surface, with that example, in terms of the implications for how you bring together an interface and interconnect different pieces of inventory into a battle network," says Dr Soare.
She added that allies can't afford to "worry about different applications and different pieces of inventory from untrusted sources being plugged and played into that battle network".
RAdm (Ret'd) Burton insists "we've got to start producing conduits that enable data to be selectively shared" but that "the prize at the end" would be a "diversity of data" offering UK allies a "significant advantage" over China.
Join Our Newsletter
