AIC4ATM — Artificial Intelligence Compliance for Air Traffic Management solutions. A modular AI Compliance Engine making AI in aviation explainable, auditable and regulation-ready.
Funded by the European Union under the SESAR Joint Undertaking · Grant Agreement No 101287448.
The rapid adoption of Artificial Intelligence in Air Traffic Management (ATM) and airport operations promises real efficiency gains — but raises critical challenges of trust, safety and regulatory compliance. The EU AI Act and related frameworks such as the EASA AI Roadmap require AI in critical infrastructure to be explainable, auditable and continuously monitored, yet dedicated aviation-specific mechanisms are still missing.
AIC4ATM closes this gap by developing a modular AI Compliance Engine that embeds ethical safeguards, operational transparency, Explainable AI and regulatory adherence for AI applications under development or approaching deployment.
The project combines a top-down analysis of European legal frameworks with bottom-up insights from ATM stakeholders — including AI deployers such as airports. Representative use cases are profiled, clustered and decomposed to pinpoint obligations, populating a graph-based Compliance Traceability Graph that links AI attributes, risks and regulatory requirements. A development roadmap guides the concept to Technology Readiness Level (TRL) 7 for real-world deployment.
AIC4ATM builds its AI Compliance Engine layer by layer — connecting technical system attributes to concrete regulatory obligations and human-centric safeguards.
Identify and cluster representative AI use cases in ATM and airport operations, decompose their components, and map European legal frameworks into a structured risk and regulatory baseline.
A graph-based repository capturing the interrelations among AI attributes, risks and regulatory requirements — enabling dynamic querying, pattern detection and automated reasoning at scale.
A risk taxonomy and compliance logic trees link identified risks to concrete obligations, supporting automated checklists and human–machine teaming.
End-to-end support from risk categorisation and automated compliance checklists to interfaces ensuring operational clarity, bias monitoring and continuous oversight.
Explainable AI, bias monitoring and continuous oversight keep humans meaningfully in control of critical decisions across the operational lifecycle.
A development roadmap and integration plan guide the concept toward real-world deployment, fostering safe, certifiable and data-driven AI innovation.
Eight work packages across two reporting periods deliver the foundational analysis, the engine design, the development roadmap, and cross-cutting management and communication.
Profiles and clusters representative ATM AI use cases, defines a system-decomposition methodology, and builds the graph-based Compliance Traceability Graph with automated querying mechanisms and baseline compliance matrices.
Lead: Sparsity (SPA)Develops the risk taxonomy and regulatory mapping, links identified risks to concrete legal obligations, and derives machine-readable compliance logic trees from risk categories.
Lead: Franck Dumortier (FD)Translates the classification logic and legal obligations into an architectural blueprint for the CE's modular six-layer architecture, with ethics-by-design at its core — delivered as initial and final concept outlines.
Lead: UPCEngages stakeholders to validate and refine the engine's functional specifications, then builds a phased development roadmap and integration plan guiding the concept toward TRL 7 deployment.
Lead: Neometsys (NMS)Administrative, contractual and financial coordination, technical quality and progress monitoring, and management of IPR, ethics, data (DMP) and the exploratory research plan.
Lead: UPCContinues overall coordination through project close: reporting, contract and consortium management, quality control, and final data management and exploratory research deliverables.
Lead: UPCProduces external communication materials (newsletters, policy briefs, briefing decks) for aviation, AI-governance and policy audiences, and coordinates with SESAR transversal activities and related Horizon Europe projects.
Lead: UPCConsolidates dissemination and exploitation through project end, finalising CDE materials and joint outreach, workshops and harmonised messaging with SESAR and EU initiatives.
Lead: UPCPublic deliverables are made openly available as they are approved. Documents marked sensitive are restricted under the terms of the Grant Agreement.
| Ref | Deliverable | WP | Due | Access | Document |
|---|---|---|---|---|---|
| D1.1 | ATM AI Use Cases Atlas | WP1 | M8 | Public | Expected · M8 |
| D1.2 | AI Use Cases System Decomposition Methodology | WP1 | M12 | Public | Expected · M12 |
| D1.3 | Automated Querying Mechanisms & Baseline Compliance Matrices | WP1 | M12 | Public | Expected · M12 |
| D2.1 | AI in ATM: Risk Mapping & Regulatory Insights | WP2 | M12 | Public | Expected · M12 |
| D2.2 | Compliance Logic & Conceptual Design Framework for ATM AI | WP2 | M12 | Sensitive | Restricted |
| D3.1 | Concept Outline — initial version | WP3 | M16 | Sensitive | Restricted |
| D3.2 | Concept Outline — final version | WP3 | M20 | Public | Expected · M20 |
| D4.1 | AIC4ATM Development Roadmap — initial version | WP4 | M18 | Public | Expected · M18 |
| D4.2 | AIC4ATM Development Roadmap — final version | WP4 | M24 | Sensitive | Restricted |
| D5.1 | Project Management Plan (PMP) | WP5 | M3 | Sensitive | Restricted |
| D5.2 | Data Management Plan (DMP) — initial version | WP5 | M6 | Public | Expected · M6 |
| D5.3 | Exploratory Research Plan (ERP) — initial version | WP5 | M6 | Public | Expected · M6 |
| D6.1 | Exploratory Research Plan (ERP) — final version | WP6 | M15 | Public | Expected · M15 |
| D6.2 | Data Management Plan (DMP) — intermediate version | WP6 | M18 | Public | Expected · M18 |
| D6.3 | Exploratory Research Report (ERR) | WP6 | M20 | Public | Expected · M20 |
| D6.4 | Data Management Plan (DMP) — final version | WP6 | M28 | Public | Expected · M28 |
| D7.1 | CDE Plan — initial version | WP7 | M3 | Public | Expected · M3 |
| D7.2 | CDE Plan — first intermediate version | WP7 | M12 | Public | Expected · M12 |
| D8.1 | CDE Plan — second intermediate version | WP8 | M20 | Public | Expected · M20 |
| D8.2 | CDE Plan — final version | WP8 | M28 | Public | Expected · M28 |
No publications yet.
Peer-reviewed articles, conference papers and other scientific outputs from AIC4ATM will be listed here as they are published, with open access provided in line with Horizon Europe requirements.
A multidisciplinary consortium of research, legal expertise, technology providers and major European airports, coordinated by the Universitat Politècnica de Catalunya.
