Dassault Flight-Tests Two AI Algorithms for the Rafale Cockpit

Dassault Aviation has flight-tested two AI algorithms aboard its Rafale fighter jet, moving the functions into an aircraft environment and positioning them as controlled, supervised support for the human crew.

Dassault developed one algorithm internally. It developed the other with Thales and cortAIx. The company says both sit within a broader effort to bring controlled and supervised AI into the cockpit, serving rather than replacing the people operating the aircraft.

The integration problem starts before the cockpit

The flight test is not simply a test of an algorithm’s output. Dassault identifies the availability and quality of operational data, whether real or simulated, as a central requirement. It also points to the need to combine domain expertise with efficient use of computing resources on an embedded platform subject to stringent constraints.

Three constraints Dassault identifies

• Data must be available and of sufficient quality, including data from real or simulated environments.
• Aviation-domain expertise must be brought together with the AI development work.
• The functions must use resources efficiently on computing hardware carried aboard the aircraft.

Support role, not an announced retrofit

The stated design places the crew at the center of the system. Dassault describes the functions as controlled and supervised cockpit AI serving human operators. That is a defined intended role, not an announcement that the Rafale has received a new operational AI retrofit.

Eligibility for a future upgrade also does not establish that either function will be installed. It does, however, mark a progression beyond development: Dassault has tested the algorithms in flight and says they have reached a level of maturity suitable for consideration in later Rafale modernization work.