By Nicholas Ortyl – Chief Engineer, Critical Infrastructure & Aviation, Leidos

Airports operate within one of the most complex and tightly regulated environments in the world. Balancing security, passenger experience, operational efficiency, and compliance is a constant challenge, particularly as passenger volumes rebound and new threat vectors emerge. Artificial intelligence (AI)-based algorithms are rapidly becoming an essential tool to help airports strengthen their security posture while improving passenger flow and resilience.

Security screening today spans multiple layers: passenger screening, cabin baggage screening, checked baggage screening, employee access control, and cargo inspection, among others. Each layer generates vast amounts of data. AI-based threat detection algorithms are suited to analyze these large, multi-attribute screening datasets in real time, supporting security efforts with more consistent and informed decision-making.

AI does not replace human expertise, it augments it. Humans remain central to the adjudication process. AI serves as a decision-support tool, helping personnel manage high data volumes, prioritize alarms, and maintain vigilance over long operational periods. This is designed to help streamline security screening, especially in high-throughput environments.

While many earlier detection systems relied on predefined image signatures or feature-based matching, modern AI systems use deep learning techniques to identify patterns and relationships within data. This approach helps detect threats regardless of shape, orientation, concealment method, or configuration. AI algorithms can improve detection performance while helping to lower false alarm rates, an important factor in maintaining checkpoint efficiency and passenger satisfaction.

People screening is also advancing. Millimeter wave systems powered by deep learning algorithms can detect both metallic and non-metallic threats concealed under clothing. Modern AI approaches focus on identifying underlying characteristics of threat items rather than relying solely on image comparison to known templates. This approach can help improve screening methodologies, regardless of gender and body type.

Computed tomography (CT) technology at checkpoints is a strong example of how AI is transforming screening operations. CT scanners generate high-resolution 3D images that provide significantly more information than legacy 2D X-ray systems. When paired with advanced algorithms, these systems can support improved detection of explosives and other prohibited items while supporting streamlined passenger processes, such as allowing electronics and approved liquids to remain in carry-on bags where regulations permit. This combination can enhance both security effectiveness and checkpoint flow.

For checked baggage, CT-based explosive detection systems (EDS) continue to evolve with adaptable algorithm libraries that can be updated as threat intelligence changes. The ability to refine detection models, subject to regulatory approval, means airports can respond more dynamically to emerging risks without requiring entirely new hardware platforms.

Beyond passenger checkpoints, AI-enhanced trace detection systems are being developed to help identify minute quantities of illicit substances, including emerging synthetic drugs. These capabilities are particularly relevant in cargo environments and customs operations, where rapid, accurate detection supports both security and law enforcement objectives.

The value of AI can increase significantly when systems are integrated rather than siloed. Open architecture platforms that connect screening technologies across the checkpoint and enterprise environment can enable cross-system analytics, allowing airports to identify operational trends, anticipate congestion, and align security resources with real-time risk indicators.

As global aviation continues to evolve, AI-based algorithms are moving from innovation to necessity.  Realizing the full benefits of AI requires disciplined governance and lifecycle management. Effective implementation requires high-quality training data, validation processes, regulatory oversight, and structured model update procedures. Emerging MLOps (machine learning operations) practices can enable continuous performance monitoring, controlled algorithm updates, and transparent validation, critical factors in regulated aviation environments.

 

Nicholas Ortyl is Chief Engineer for Commercial Infrastructure at Leidos, leading development of advanced security and detection technologies for airports and critical infrastructure. With more than 20 years of experience across aerospace, autonomous systems, and government security programs, he focuses on integrating artificial intelligence, advanced imaging, and data-driven solutions to enhance aviation security effectiveness, operational efficiency, and regulatory compliance.

DISCLAIMER

This article was provided by a third party and, as such, the views expressed therein and/or presented are their own and may not represent or reflect the views of Airports Council International-North America (ACI-NA), its management, Board, or members. Readers should not act on the basis of any information contained in the blog without referring to applicable laws and regulations and/or without appropriate professional advice.