Nato allies need to speed up AI defence co-operation

As Russia intensifies cyber hostilities and China weaponises artificial intelligence, joining forces in the field of high-tech warfare will feature high on the list of topics discussed by Nato allies at a summit next week. But the transatlantic alliance’s 30 members will need to move fast if they aim to make up lost ground.

Nato is proposing a new tech innovation centre bringing together military personnel with industry to foster digital defence start-ups. Some of these might be financed by a separate initiative, also set to be debated: a venture capital fund for innovation which member states could choose to opt in to.

The efforts are belated, as Nato secretary-general Jens Stoltenberg himself acknowledged. “For decades, Nato allies have been leading when it comes to technology, but that’s not obvious any more,” he told the Financial Times in an interview last week. “We see China especially investing heavily in new, disruptive technologies like artificial intelligence, autonomous systems, big data, and they implement them into new advanced weapon systems, drones, submarines, aircraft and so on.”

He is not the first to sound the alarm. Eric Schmidt, the former Google chief executive who now chairs the US’s National Security Commission on AI, warned earlier this year that Beijing was planning to undermine conventional military forces by “leapfrogging” to new technologies. The commission’s report, published in March, raised concerns that China would use AI for “reconnaissance, electromagnetic countermeasures and co-ordinated firepower strikes”.

Part of the problem is that western defence institutions have been slow to recognise the potential of innovation beyond their own industry.

For decades, a lot of technological development would happen within the defence sector . . . and then shared with the civilian sector. Now, it goes the other way around

“For decades, a lot of technological development would happen within the defence sector — the internet, nuclear, GPS, all of that was developed by the defence industry and then shared with the civilian sector,” Stoltenberg said. “Now, it goes the other way around. It’s a civilian sector which is leading in the development of artificial intelligence, quantum computing, and many of the new disruptive technologies.”

Some Nato members are ahead of others. The US and France have published military AI strategies, while the UK announced this year that it is to establish a centre for defence AI. For the first time, Britain’s intelligence agency, MI6, is recruiting from the private sector for a new head of its “Q” branch — the technical lab made famous in the James Bond films.

Establishing a new Nato hub — known as an “accelerator” — in which tech companies and members of the armed forces can experiment with new ideas has advantages, according to Professor Fiona Murray, co-director of MIT’s innovation initiative.

Start-ups and investors do not always have the time to tackle defence challenges when solutions are “hard to test, markets are fragmented and procurement is slow”, Murray said.

Working together would create a wider market for new products and enhance collective security, she noted. It was “not enough” for countries to be handling this individually, she said.

The US has started marshalling allies on the policy implications of using new technology. The Pentagon’s “AI Partnership for Defense”, comprising 13 countries (including Nato members Canada, Denmark, Estonia, the UK, France and Norway) met virtually for the first time last year to agree joint military standards on AI. Schmidt’s commission has called on the Five Eyes intelligence-sharing alliance (the US, UK, Canada, Australia and New Zealand) to work more closely on developing AI systems.

Ulrike Franke, an expert in military technology at the European Council on Foreign Relations, argues that Nato’s tech centre will be most effective if it prioritises systems designed to facilitate joint military operations. The alliance should look at areas such as AI-enabled command and control, she said, which could give members a unified picture of the battlefield across multiple regions, using intelligent data analysis to sift information.

Franke said that in the vast arena spanning drones to quantum computing, there was a temptation to cover too much. “It makes massive sense for Nato to look more at this [technology]”, she said. “The question is, what exactly are they focusing on? There’s a danger of Nato spreading itself too thin.”