日本語 
Dassault Aviation has tested two artificial intelligence algorithms on the Rafale fighter jet. This happened during flight tests in September. The tests took place on September 22. The company said the goal is to bring supervised AI into aircraft. The AI is not meant to fly the plane on its own. Instead it is meant to help the pilot.
One of the AI algorithms was created by Dassault Aviation engineers. The other was developed with Thales. Thales contributed its cortAIx AI accelerator. Both algorithms were tested in flight conditions.
These tests are part of a plan. Dassault is working to add controlled and supervised AI into the Rafale cockpit. The technology is designed to support pilots. It is not meant to replace them.
Dassault said both AI functions are now mature enough. They are ready to be considered for upgrades to the Rafale. The company sees this as a step forward. It shows progress, toward more capable fighter jets.
AI Moves From the Lab to the Aircraft
The flight testing is significant because military AI must operate under very different conditions from conventional enterprise software.
An aircraft’s onboard computing environment is constrained by power, weight, processing capacity and the need for reliable operation in demanding conditions. AI systems also have to work with operational data that can be limited, incomplete or difficult to reproduce.
Dassault identified three major challenges: ensuring the availability and quality of real or simulated operational data, combining AI with domain expertise, and optimizing computing resources on an embedded platform with strict constraints.
These requirements make embedded AI an important engineering problem. Instead of relying entirely on remote cloud computing, algorithms used in aircraft need to process information locally and deliver results within the operational environment.
Sovereign AI Becomes a Defense Priority
The development also reflects growing interest in sovereign AI capabilities in the defense sector.
Dassault’s strategy calls for secure, sovereign and supervised AI in combat systems. The company has established partnerships with France’s Ministry of Defense AI Agency, Thales through its cortAIx initiative and Harmattan AI to accelerate supervised autonomy and AI integration into aviation systems.
The distinction between sovereign and externally supplied AI can be important for defense organizations because military systems involve sensitive data, specialized operational knowledge and national-security requirements.
Developing critical AI capabilities domestically can also give defense manufacturers greater control over software updates, system integration and the underlying technology stack.
Rafale’s AI Roadmap
The new AI tests form part of the Rafale’s continuing technology evolution. Dassault is developing the F5 standard, which is intended to expand the aircraft’s capabilities and support collaborative combat involving unmanned systems.
The company has already demonstrated another element of this direction. In July, Dassault Aviation and Harmattan AI conducted a flight involving a Rafale F4 and an unmanned aerial system carrying the NAMIB electronic-warfare payload. NAMIB detected and geolocated electromagnetic emissions from an air-defense radar, with the information transmitted to the Rafale during the test.
Together, these developments show how AI is increasingly being integrated into the aircraft, sensors and surrounding systems rather than treated as a standalone software capability.
Implications for Japans Defense Technology Sector
The development matters to Japan because the country keeps investing in AI, autonomous systems, sensors and advanced defense technologies.
Japans defense technology ecosystem includes electronics, communications, robotics and industrial technology companies. These abilities can help in areas such as edge computing, radar processing, image recognition, electronic warfare, autonomous systems and secure communications.
Embedded AI could be especially useful for platforms that work where reliable connectivity to computers cannot be expected. Aircraft, ships, unmanned vehicles and remote sensors may need processing to read data and help human operators in real time.
The technology also links to Japans goal of strengthening domestic and allied supply chains for advanced computing and AI.
In July 2026 Japan and India also agreed to deepen cooperation in AI including AI governance, safety, cybersecurity and trustworthy AI supply chains. Although that agreement is not a defense program it shows the broader regional focus on resilient and trusted AI capabilities.
Human Oversight Remains
Dassaults description of the technology is notable because it stresses controlled and supervised AI. The goal is to help the crew not to remove human involvement from the system.
That approach matters beyond aviation. As AI moves into safety- environments developers must address reliability, explainability, cybersecurity, testing and human oversight together with raw model performance.
For technology suppliers this creates chances, around AI accelerators, sensors, simulation environments, cybersecurity and verification tools. It also makes testing AI systems against difficult scenarios more important before they are taken for operational deployment.
From Fighter Aircraft to a Broader AI Ecosystem
Dassaults Rafale tests show a change in aerospace. The change is toward platforms that are defined by software and that use AI. Future aircraft are expected to bring onboard computing, sensors, communications and autonomous abilities. At the time human control will still be kept.
Also Read: Liberaware Selected for Tokyo’s SusHi Tech Initiative
For Japans technology industry this development shines a light on chances. Embedded AI, edge computing, aerospace electronics and autonomous systems all become possible. The result also shows why defense AI is not a matter of building a model. It is also a challenge of engineering and data.
The newest Rafale tests do not say that AI capabilities are already working on every plane. Dassault said the functions that were tested have reached a maturity level that lets them qualify for upgrades. This means development and more tests will still be needed before AI can be put into service.
As military platforms grow more software driven the skill to build AI that is secure, reliable and locally controlled could grow more and more important, in the global aerospace technology industry.
English