2025 Agenda

Abstract:
As automation and AI become central to the future NAS, traditional certification and verification approaches are increasingly mismatched to reality. Existing assurance processes are largely linear, document-driven, and designed for deterministic systems with infrequent updates, an approach that does not scale to adaptive algorithms, data-driven functions, and continuously evolving software. At the same time, safety requirements remain uncompromising.
This session focuses on how verification, validation, and certification practices must evolve to support highly automated and AI-enabled systems that learn, adapt, and interact with other intelligent services, generating behaviors at the system-of-systems level.
Discussion will address:
• Limitations of current certification models
Where conventional means of compliance fall short when applied to learning systems and adaptive autonomy.
• Verification and Validation (V&V) for AI and autonomy
Methods for assessing systems whose behavior is influenced by training data, operational context, and interaction with other agents.
• Run-time assurance architectures
Safety monitors that bound AI behavior, detect unsafe states, and intervene—reducing the need for full recertification after every model or data update.
• Evidence generation in digital environments
he role of digital twins, large-scale simulation, and scenario generation in building regulatory confidence and enabling operational approval.
• Equivalent processes and alternative means of compliance
Regulatory pathways that support performance-based evidence, defined operational envelopes, and structured risk cases as complements to prescriptive technology requirements.
• Assurance at the system-of-systems level
Approaches for evaluating emergent safety behavior across interacting ATM, UTM, AAM, and ETM systems spanning aircraft, ground infrastructure, and data services.
The session will emphasize practical mechanisms to accelerate certification without compromising safety, including modular approval approaches, incremental authorization, and continuous assurance based on operational data.
Participants will gain a clear understanding of the technical, regulatory, and organizational changes required to safely certify an autonomous, data-driven NAS, and the respective roles of regulators, ANSPs, researchers, and industry in achieving that outcome.