The European Union Aviation Safety Agency is moving its artificial intelligence program toward aviation-specific regulations as developers demonstrate systems for cockpit decision-making, safety analysis and certification during EASA AI Days in Cologne, Germany. EASA’s rulemaking manager for AI is providing an update Thursday on RMT.0742, the agency’s effort to establish requirements for the use of AI across regulated aviation activities, according to the conference event agenda.
Rules Move Toward Aviation Applications
EASA published its first proposed AI regulations last November. The proposal establishes general standards covering AI assurance, human factors and ethics for systems that assist people or work cooperatively with them. Public comments closed in March, and EASA’s rulemaking system currently says the proposal is awaiting responses to comments. The agency has said a second proposal will apply that general framework to individual aviation domains, and its formal rulemaking task includes adapting the requirements for specific regulated areas and other operations, including aerodromes.
“These applications open the way to novel types of operations in which the human end user may be either remotely present, or not present during the operation,” EASA said when it released the latest version of its concept paper in June.
That document expanded the agency’s work to include reinforcement learning, symbolic AI and Level 3 systems, which EASA classifies as advanced automation.
Cockpit AI Among Systems Tested
The agency is also using the conference to examine 10 AI systems selected from 50 submissions for its first AI Olympics. The finalists include Kairós Aeroespacial’s FOR-DEC AI, a cockpit system designed to help crews organize operational information, compare options and identify uncertainty during decision-making.
Other finalists include an AI system for analyzing aviation safety reports and Xcert AI, which is designed to assist engineers working through certification and regulatory-compliance requirements. EASA published the finalist list Wednesday, and conference attendees are voting to select gold, silver and bronze projects.
“The pilot remains the decision-maker,” Kairós said in describing its cockpit system. The company says the software is intended to make its reasoning traceable to the flight crew rather than independently make operational decisions.
EASA’s remaining sessions Thursday include research into large language models and reinforcement learning, followed by a panel examining the benefits and risks of advanced automation.
0 replies