Continuing our series on use cases of AI in government, we cover how the TSA is using AI to improve service.
TSA deploys AI within its Checkpoint Property Screening System (CPSS) to automatically detect non-explosive prohibited items in carry-on baggage using Computed Tomography (CT) scanners, which generate 3D images for precise threat identification. This improves government efficiency by accelerating passenger throughput at security checkpoints, minimizing manual inspections, and ensuring consistent threat detection, ultimately reducing wait times and enhancing overall airport security operations.
Evaluation of Technology: The system employs machine learning algorithms for image segmentation and object recognition, producing 3D bounding boxes around potential threats like firearms or sharp objects on X-ray scans. Strengths include superior accuracy over traditional 2D X-ray systems due to volumetric analysis, automation that lightens the workload on security officers (allowing focus on high-priority alerts), and adaptability through model retraining for emerging threats. Limitations encompass reliance on high-quality training datasets to avoid biases or misses on atypical items, computational intensity requiring robust hardware, and the necessity for human validation to handle edge cases or false positives. Overall, it’s highly effective for high-volume screening environments, boosting detection rates while maintaining safety, but benefits from ongoing algorithm updates and integration with open architecture standards for interoperability.
Financial Resources: The CPSS program is funded through TSA’s Capital Investment Plan, with $144.7 million allocated for FY2025 and a total of $731.7 million projected from FY2025 to FY2029 for procurement and deployment of CT units incorporating AI capabilities. Additional potential funding of up to $250 million annually could accelerate full operational rollout by supporting R&D for enhanced detection standards.