Article 108: Amendment to Regulation (EU) 2018/1139

Article 108 integrates the AI Act into Regulation (EU) 2018/1139 — commonly known as the EASA Basic Regulation — which is the cornerstone of EU civil aviation safety law. This Regulation establishes the European Union Aviation Safety Agency (EASA), sets out essential requirements for airworthiness, air operations, aircrew, air traffic management (ATM), and aerodromes, and defines the certification and oversight framework for all civil aviation in the EU.

The amendment means that where AI systems are used in any aspect of aviation covered by the EASA Regulation — from aircraft design and certification to air traffic management, predictive maintenance, UAS/drone operations, and ground handling — the AI Act requirements must be considered as part of EASA's regulatory processes. This follows the Article 8 integration model.

This is one of the most consequential amendments because aviation has some of the most stringent safety requirements of any industry, and AI is being adopted rapidly across the sector. The amendment ensures that EASA's established safety framework — including its risk-based approach, design assurance levels, and continuing airworthiness requirements — is enhanced to address AI-specific risks.

• Article 8 — Integration with sector legislation: AI Act requirements are assessed through existing aviation certification and approval processes. Article 108 brings Regulation 2018/1139 into this framework.
• Annex I — Union harmonisation legislation: the EASA Regulation is listed, making aviation AI systems subject to the integrated conformity assessment regime.
• Article 9 — Risk management: aviation AI risk management must meet the highest standards, consistent with the sector's safety culture and existing risk management frameworks (e.g., EASA's management system requirements).
• Article 14 — Human oversight: pilots and air traffic controllers must retain meaningful decision authority over AI-assisted operations — consistent with established crew resource management (CRM) and controller responsibility principles.
• Article 102 — Aviation security amendment: Article 102 covers aviation security equipment under Regulation 300/2008; Article 108 covers aviation safety under the EASA framework. Together they address the full aviation regulatory landscape.
• Article 113 — Entry into force and application dates: confirms the timeline.

For aircraft and component manufacturers (Design Organisations): - Map AI systems across your product portfolio: autopilot enhancements, flight envelope protection, engine health monitoring, AI-assisted diagnostics, and any ML-based design or simulation tools used in the certification process. - Prepare AI Act compliance evidence as part of your type certificate (TC) or supplemental type certificate (STC) applications. EASA will expect to see AI-specific documentation alongside existing certification data packages. - Monitor EASA's guidance development — EASA has been actively working on AI/ML certification frameworks (e.g., the EASA AI Roadmap and Concept Paper on AI). These will inform how Article 108 is implemented in practice.

For airlines and operators: - Assess AI tools used in operations: dispatch systems, predictive maintenance platforms, pilot tablet applications, and fuel optimisation algorithms. Even if these are not part of the aircraft's type certificate, they may fall under operational approvals or management system requirements. - Ensure that flight crew are trained on AI system capabilities and limitations — this supports both the AI Act's human oversight requirements and existing CRM/training obligations.

For ATM/ANS providers: - Identify AI components in ATM systems: arrival/departure management, conflict detection and resolution, trajectory prediction, and workload management tools. - Implement AI Act requirements within the existing ATM/ANS management system and change management processes, coordinated with EASA and national supervisory authorities.

For UAS/drone operators and manufacturers: - AI is fundamental to most UAS operations (autonomous flight, sense-and-avoid, payload management). Ensure AI Act compliance is integrated into EASA UAS certification (for certified category) or operational authorisation processes.