Posted by 
It has taken a while, but Europe has recently recognized that for pressing defence issues a software solution is very often state of the art. Boosting defence today is often less a matter of spending ever-increasing amounts on expensive hardware: military planners now recognize that software can drastically augment the performance of new platforms as well as retrofit legacy equipment, while being cost-effective compared to large platform procurements. It can also rejuvenate the combat power of European forces limited in mass.
This shift in understanding has boosted our ability to safeguard European sovereignty in the face of technological progress by adversarial actors. In the past 12 months, the first large contracts have been awarded to pure software defence companies, and while more work remains to be done, European governments have shown they can adapt, act and achieve state-of-the-art defence.
Europe has also thankfully absorbed another insight: space has become a new frontier in security. Elon Musk’s privately-funded Starlink satellite constellation has been the communications backbone of Ukrainian forces for much of the conflict with Russia. The success and positive effect of this capability speak to a larger realisation: sovereign communication and satellite capabilities are core to the security of European nations. The Continent’s armed forces must be able to rely on SatCom connectivity to exchange battle intelligence and achieve information superiority. Consider land forces looking to speed up their decisions chains, fleets of ships, submarines and helicopters operating together, or, further afield, teams of manned and unmanned aerial combat vehicles seeking to dominate the airspace — reliable connectivity is and will remain key to success on the battlefield.
For all of the promise of IRIS², it lacks one crucial component: artificial intelligence
Enter IRIS²: the Infrastructure for Resilience, Interconnectivity and Security by Satellite, the European Union’s satellite constellation initiative poised to restore the Continent’s sovereign capability in space communication. It is set up as a potential game-changer: a versatile system serving civilian, governmental, and military needs; robust security measures; and a forward-thinking design that anticipates future upgrades. In essence, IRIS² is on track to be Europe’s solution to provide robust communication around the globe. However, as things stand, IRIS² runs the risk of being outdated before it even launches.
For all of the promise of IRIS², it lacks one crucial component: artificial intelligence (AI). The reason AI is not yet on board is simple: When the programme was originally launched, AI was still seen as a somewhat futuristic technology with its potential not yet fully understood. However, over the past two or three years, the world has learned that AI has matured and is ready for deployment practically everywhere.
Moving forward, we believe that AI needs to be at the core of every new satellite constellation to reliably function and stay relevant for years to come. There are two reasons for this:
First, modern satellites capture vast amounts of data, yet are bandwidth-limited in transmitting this data down to ground stations. Only AI on the satellite itself makes this data manageable, pre-processing and analysing it in space before selectively sending down insights and data.
As an example, instead of sending gigabytes of sensor data to a human operator on the ground, an AI-enabled satellite would first analyse the data, look for patterns of interest, and then send down the distilled information (for example, “hostile radio frequency emitter at location x/y/z”). The resulting speed of transmission and efficient use of bandwidth significantly reduces reaction time for crucial analysis. We must assume that competitive countries like China have designed this key capability into their constellations — Europe’s IRIS² must not launch without it.
If IRIS² lacks AI, its satellites will be unprotected against adversarial attacks
Second, satellite constellations are increasingly becoming the target of cyberattacks, and their transmissions are subject to ‘smart jamming’. To counter this emerging threat, satellites require AI to detect sophisticated attacks and to counter them in an adaptive manner — AI essentially provides an immune system to fend off shape-shifting intruders.
Adversaries will attempt to jam, or interfere with, the signal the satellite is sending to earth. AI-based anti-jamming technology automatically counteracts even the most complicated jamming attacks — and can learn from novel patterns to react even faster over time. If IRIS² lacks these AI countermeasures, its satellites will be unprotected against adversarial attacks.
Will the EU and its contractors be able to adapt the programme?
AI will be the defining technology of the coming decades, and its benefits to Europe’s new satellite constellation are specific and fundamental. The standards and capabilities for IRIS² are being decided in the coming weeks and months. We urge policymakers involved to not treat AI as a ‘nice-to-have’ for IRIS², but instead make it a core requirement for the system. Only then can IRIS² become — and remain — truly state of the art.
General Denis Mercier is the former Supreme Allied Commander Transformation of the North Atlantic Treaty Organization and former Chief of Staff of the French Air Force. He is a member of the board of directors of the defence AI company Helsing.
Marc Fontaine is président of the defence AI company Helsing France. Before Helsing he served as chief digital transformation officer at Airbus.
