Artificial intelligence seems to have established itself firmly in the defense industry. Last month at Eurosatory 2024 in Paris, there was a wide range of AI-enabled solutions on offer from both startups and major defense companies. But this widespread adoption comes with its own challenges, including issues of trust.
On June 19, Safran announced the launch of its Advanced Cognitive Engine (ACE) system, which aims to integrate AI capabilities into all Safran Electronics & Defense products to enhance situational awareness, support decision-making and reduce the cognitive load for troops in the field.
At this point in the development and integration of AI into the defense and naval world, decision support is undoubtedly one of AI's primary benefits. A key driver for this is the huge disconnect between the rapid increase in sensor data collected in operational areas and the challenges militaries face in recruiting and retaining talent.
Related article
How Artificial Intelligence Could Threaten Military Readiness
U.S. to boost funding for artificial intelligence capabilities in fiscal year 2024
How AI Tools Can Reduce Fuel Use on Offshore Platforms
“Autonomous systems are changing the way we collect data, but they also create significant personnel challenges, as currently each drone or robot requires an operator,” said Arnaud Valli, communications director at Command AI. ShepardDuring Eurosatory 2024.
“Given the challenges all militaries face in recruiting, training and retaining personnel, AI could be a real game-changer in reducing the cognitive load that all this data places on personnel.”
Against this backdrop, French start-up Command AI has developed two AI-enabled systems to improve operational readiness by supporting mission planning and command and facilitating training.
By integrating multiple AI modules into an off-the-shelf computer, multiple scenarios and solutions can be easily generated based on collected and processed data. Valli and company founder Loïc Moujoll have been preparing two modules for testing in November following a call for interest from the French Defense Innovation Agency.
Additional wargaming-focused modules are also being developed, and the Command AI system can be used by both Army and Navy.
Similarly, MBDA discussed the integration of its new AI-enabled decision support kit, Ground Warden, into its Akeron family of weapon systems.
“The aim is to enable these systems to be used beyond visual line of sight,” Matthieu Crowley, the company's head of business development for battlefield systems, told reporters at the show.
Connected to the firing range, the Ground Warden uses information about the target to guide the Akeron operator to ensure the elimination of stationary threats.
“AI will be used to continuously track threats in all weather conditions,” Crowley explained.
Ground Warden can also be used in conjunction with autonomous systems to assist operators in effectively eliminating moving targets.
Akeron MP is now enhanced with a unique Naval Mode, an algorithm designed to improve seeker stabilization and image processing.
A question of trust for AI-powered naval platforms
While AI is key in enabling autonomy and reducing the cognitive load on operators, whether military forces can fully trust AI-enabled systems remains a major question and challenge.
MBDA has been very open about the subject, explaining that AI is currently used to support decision-making, but full autonomy is still a long way off as many questions remain: Who will provide the data for training the AI, who will carry out that training, how will it be certified, etc.
Thales, for example, is developing contracts designed to enable operators to define the rules and limits of a system's autonomy and deploy it with confidence, which is what its SwarmMaster solution enables.
At Eurosatory, ShepardWe also met with Numalis, a French startup that develops tools for developing reliable and explainable AI, including its Saimple product, which is designed to help AI developers validate their algorithms according to the ISO/IEC 24029 standard.
“The concept of Saimple is to validate AI algorithms at every stage of the development process to make sure they continue to function correctly despite potential disruptions,” said Jack Mojsilovic, chief communications officer at Numalis. ShepardThese disruptions could be bad weather or other factors that could affect the performance of an AI-enabled system.
To achieve this, Numalis uses two metrics: Dominance, which is a metric to determine when a decision the AI makes can no longer be trusted, based on the AI's possible operating space, and Relevance, which is a metric based on visual aids that show the characteristics of the object the AI used to make its classification.
“Solutions like Saimple aim to shorten the development and validation cycles of AI algorithms, making it possible to check their robustness and explainability at every stage of development,” said Arnaud Ioaralen, CEO of Numalis. ShepardEventually it will support AI authentication.
Numalis is also collaborating with companies such as Safran and MBDA to develop more robust and explainable AI.
There are other efforts in the defense industry to develop algorithms that are transparent, explainable, and trustworthy, and while they are moving forward quickly, the road ahead is more toward mid-term goals than short-term goals. Until we get to that point, AI will remain a decision support tool for human-involved systems across all military forces.
This analysis article was published in the June issue of the Decisive Edge Naval Newsletter. To receive regular updates from Alix Valenti and our team of defense experts, Decisive Edge sign-up page.
Defense Insight Related Equipment
Acheron MP
