France has launched the Man-Machine Teaming study into applying artificial intelligence to combat aircraft, part of a €1.5bn investment strategy aimed at closing the technology gap with the US and China. Julian Turner reports on the global race for supremacy in the sphere of military AI.

On 16 March, French Armed Forces Minister Florence Parly announced a dramatic increase in the country’s annual spend on artificial intelligence (AI) to €100m – a total of €1.5bn ($1.83bn) in public investment by 2022 – as part of France’s innovation drive to develop future weapon systems.

The announcement coincided with the launch of the government-backed Man-Machine Teaming (MMT) study for applying AI to combat aircraft – part of a roadmap to explore AI for armaments – and the creation of a defence innovation agency open to civil, startup and European investment.

Germany’s Minister for Education and Research Anja Karliczek has already stated that Germany and France could link up their data centres and collaborate on bilateral research programmes, while President Macron said that a portion of the funds would bankroll Franco-German research projects.

Less than two weeks later, a parliamentary report, 'Giving meaning to artificial intelligence, for a national and European strategy', confirmed that AI is now a central policy tenet, predicting that the mastery of such technologies will become a state necessity to "ensure security missions, maintain the ascendancy against our potential adversaries" and "maintain our position in relation to allies".

France’s €1.5bn investment is all the more impressive given that the European Commission has pledged to boost spending on AI in the whole of Europe to the same figure by 2020, an increase of around 70%.

The MMT study for the Direction Générale de l’Armement (DGA) is run by Dassault and Thales, and will look at the feasibility of flying fighter jets and drones together to evade air defence systems.

Roughly half the French annual AI budget will fund studies, with €10m a year earmarked to test and integrate existing AI technology. A total of 50 AI specialists will be recruited by 2022 to staff the DGA procurement office, and the armed forces ministry will also monitor the development of civil AI.

The flurry of military AI research and development (R&D) in France encompasses start-ups, SMEs and laboratories specialised in AI, robotics and new man/machine interfaces, and focuses on the concept of a ‘cognitive air system’ concept, which aims to increase autonomy onboard aircraft.

Roughly half the French annual AI budget will fund studies.

According to Air & Cosmos International, the MMT study goals include: defining future cockpits and independent systems; improving man/machine teaming technologies (particularly decision-making autonomy and machine learning) within the cognitive air system; and improving smart/learning sensor concepts and technologies.

AI algorithms already calculate missile trajectories, perform transcriptions and translate foreign languages. However, the DGA has identified a range of potential applications including: electronic warfare and intelligence gathering; automatic image recognition; collaborative combat on land and in the air; autonomous navigation robots; cybersecurity; predictive equipment maintenance; anti-mine warfare; and decision and command support.

Future military aircraft will be fitted with advanced sensors that gather a multitude of data; the MMT study will focus on how best to process and merge this in real time as well as integrate it with legacy data and other information sources in the combat cloud, reports Defense News.

The strategy is in part an attempt by the Macron administration to keep pace with China – which plans to become "the world's primary AI innovation centre" by 2030 – and stem the brain drain of computer engineers and mathematicians from France to US technology companies in Silicon Valley.

Asia and the US are each investing at least three times more in AI than Europe. In 2016, European private investments totalled €2.4-€3.2bn, compared to almost €10bn in Asia and €18bn in the US.

“There is a global arms race going on and the price to compete keeps going up and up,” says Jonathan Schaeffer, a computing scientist at the University of Alberta in Edmonton, Canada.

“It is hard to envision the future, but with the potential that AI has, a major investment today might be insufficient tomorrow.”

To match the flexibility of universities in other countries, Macron is seeking to increase the time that publicly funded scientists are allowed to spend working in private companies, from 20% to 50%.

“There is a global arms race going on and the price to compete keeps going up and up.”

Speaking at the AI for Humanity conference in Paris, Macron also touched on the ethical dimension to France’s €1.5bn investment in AI, saying that turning the country into an AI leader would allow it to use AI for the public good and ensure that a “Promethean” promise doesn’t become a “dystopia”.

The French President wants to ensure that AI algorithms are controlled for the greater good in order to avoid the “opaque privatisation of AI or its potentially despotic usage” by foreign governments.

This could entail the formation of a group similar to the Intergovernmental Panel on Climate Change (IPCC), which comprises thousands of volunteer scientists who review climate science literature.

Compared with China and the US, France is still playing catch-up when it comes to developing military applications for AI.

Writing in the Indian Defence Review, Deepak Kumar Gupta cited China’s recent advances in swarm intelligence, which involves autonomous cooperative behaviour among masses of distributed robots.

In June 2017, state-owned defence conglomerate China Electronics Technology Group Corporation successfully flight-tested a swarm of 119 drones, a new record.

“In a conflict, the PLA [People’s Liberation Army] could use swarms to cheaply target high-value US weapons platforms, such as aircraft carriers,” wrote Gupta.

“The US Government is spending billions of dollars preparing for the next stage in warfare that it believes will be defined by advances in AI,” he continued.

“Concepts like motherships of drones releasing little baby drones from the air and the sea, infantrymen and women sporting exoskeletons and wearable electronics loaded up with combat apps, and lone mission commanders directing swarms of unmanned vessels to carry out operations are already being tested at MIT’s Computer Science and AI Laboratory.”

“The PLA [People’s Liberation Army] could use swarms to cheaply target high-value US weapons platforms.”

AI’s central role in 21^st century defence and security has also been recognised by the UK. On 21 May, Defence Secretary Gavin Williamson launched AI Lab – a single flagship for AI, machine learning and data science at the Defence Science and Technology Laboratory (Dstl) in Porton Down, Salisbury.

Dstl already delivers more than £20m of AI research, a figure that is forecast to grow significantly. Williamson unveiled AI Lab during the first ever joint US-UK Defence Innovation Board meeting.

AI Lab will engage in high-level research on everything from autonomous vehicles, intelligent systems and countering fake news to using information to deter and de-escalate conflicts, enhance computer network defences and improve decision aids for commanders.

Across the English Channel, things are also moving apace. Technology from France’s MMT study will be applied to all future manned and unmanned combat aircraft systems, with the first AI applications expected in 2025 – followed by widespread dissemination as soon as 2030.