2025 Agenda

Abstract:
Over the past decade, artificial intelligence in aviation has been deployed primarily as advisory capability, improving trajectory prediction, anomaly detection, and decision support for human operators. The next decade requires a more fundamental transition: from AI that advises humans to AI that acts as an operational participant. In the future NAS, tool-centric applications will evolve into system-level AI, in which multiple intelligent services coordinate to manage complex operations across ATM, UTM, AAM, and ETM domains. Autonomous digital agents will negotiate airspace access, reconcile conflicting objectives, and coordinate traffic flows, frequently at speeds beyond direct human control.
Discussion will address:
• Autonomous negotiation among aircraft and airspace services
Multi-agent systems supporting separation management, dynamic sequencing, and cross-domain prioritization.
• Global data reconciliation and common APIs
How heterogeneous systems such as ATC decision support tools, airline operations systems, UTM service providers, and weather services share state information through standardized semantics while preventing new failure modes.
• Persistent digital twin operations
Continuous, high-fidelity digital representations of the NAS used for prediction, safety monitoring, what-if analysis, autonomous negotiation, and recovery planning.
• Continuous safety assurance and learning systems
AI that does not just perform a function but continuously evaluates its own performance, detects drift, and adapts within certified bounds.
• Handling off-nominal and degraded operations
How system-level AI responds to unexpected conditions such as:
− rapidly evolving convective weather
− GNSS interference or denial
− demand surges and metering constraints
− mixed-equipage environments and partial automation failure.
Panelists will examine architectures that support resilient, adaptive autonomous behavior while preserving operator authority, traceability, and auditability. The session will explicitly address where full autonomy is appropriate, where human-on-the-loop oversight remains essential, and how decision authority transitions during contingency operations.
Attendees will gain a clear understanding of how AI functions not merely as a system add-on, but as an organizing principle for the future NAS, one that enables the system to maintain safety, capacity, and efficiency during both normal operations and major disruptions, at national and global scales.