Defense analysts in operations centres face significant challenges in extracting relevant information from the huge volumes of data generated by multiple sources such as video and audio streams, websites, Twitter feeds, satellite imagery, social media and telephone conversations. Real-time data analytics will enable them to develop advanced military strategies with greater efficiency. Thales, an expert in artificial intelligence, plans to implement NukkAI’s solution in a number of its military data processing programmes. When operators are swamped by information, the solution will use real-time data exploitation and fusion methods to automatically review the knowledge available so that analysts can focus on elements of interest.
Jean-Baptiste Fantun, co-founder and CEO of NukkAI, said: “There are multiple use cases for our solution in cybersecurity, education, industry, banking and insurance and in any other area where data from multiple sources needs to be combined, where outcomes need to be explainable and where human operators need to retain control at all times.”
David Sadek, Vice President for Research, Technology & Innovation, who is in charge of AI projects at Thales, said: “This pilot project represents a real technological milestone for military applications of AI, enabling analysts to focus on the tasks where human beings can provide the most added value. The partnership with NukkAI is fully in line with our Thales TrUE AI approach, which favours the use of AI solutions that are trusted, safe, secure, explainable and responsible.”
The explainable AI technology developed by NukkAI was put to the test during the world bridge tournament in March 2022, successfully beating eight human world champions at the well-known card game. The demonstration confirmed the potential of this approach for the development of innovative applications in various sectors of industry. NukkAI’s solution relies on the use of hybrid, explainable, collaborative and energy-efficient artificial intelligence methods, unlike the “black box” algorithms in widespread use today, which lack transparency and raise issues of human-machine interaction and high energy consumption.
NukkAI’s AI uses a combination of modules relying on different paradigms of artificial intelligence, symbolic AI and digital intelligence. While “black box” AI is not designed for human-machine interaction, NukkAI’s solution can interact with humans and explain the reasons behind its choices and decisions. And it is the human who makes the decision, following suggestions made by the machine. Developed with the support of the CNRS, which provided access to the Jean Zay – the most powerful supercomputer in France – the NukkAI solution that won the bridge tournament in March 2022 consumed 200,000 times less energy than the AI used to beat the world Go champion. For the bridge tournament, NukkAI developed a tool to analyse game play and explain the strategy used, making it possible to transfer the robot’s skills to the human or show human how their strategy was inferior to that of the robot.