EASA AI regulations for aviation

EASA’s first-ever AI rules aim to bring safety, transparency and human oversight to AI in EU aviation.

On November 14, 2025 the European Union Aviation Safety Agency (EASA) published a regulatory proposal targeting AI in aviation — framed throughout as EASA AI regulations. The draft sets out baseline requirements for safety, certification pathways and operational monitoring for AI-powered systems used in aircraft systems, air traffic management (ATC) and aircraft maintenance across Europe.

The proposal is positioned as a harmonising step for Western aviation, addressing risk, explainability and how automated decisions are audited. EASA says the rules will apply to development and in-service monitoring, with measures to make AI decision-making more transparent to crews, controllers and maintenance teams. A public consultation is expected to open in early 2026.

EASA AI regulations: what’s in the proposal

Details are still high level in the published text, but the agenda is clear: certification, continuous oversight and human-in-the-loop safeguards. The draft asks manufacturers and service providers to demonstrate how AI models are validated for safety, how outputs will be monitored in real time, and how human operators will retain meaningful control. The agency also emphasises post-market monitoring to catch degraded performance or unexpected behaviour once systems are live.

Certification pathways for AI systems (design, testing and approval)
Requirements for post-market monitoring and reporting
Human oversight, training and fallback procedures
Demand for transparency and explainability in AI decision-making

Industry stakeholders should prepare for a consultation phase that will allow input from airlines, manufacturers and air navigation service providers. While the proposal does not single out aircraft types or models, its scope—aircraft systems, ATC and maintenance—means both OEMs (original equipment manufacturers) and airlines will need to assess how AI integration aligns with new expectations for safety and traceability.

Next steps: expect technical working groups and a public comment period in early 2026, after which EASA will refine the draft before formal rulemaking. For now, the headline is straightforward: EASA AI regulations are designed to bring AI into aviation under a safety-first, transparent framework.