Lauretta AI announced a collaboration with the Transportation Security Administration to test and evaluate Delos, its behavioral video-analytics system. The announced work focused on improving detect-and-track capabilities through structured evaluation and feedback rather than presenting the technology as a finished replacement for security personnel.

What Lauretta announced

According to the company’s original press release, TSA’s Multimodal Passenger Assessment and Consequence Evaluation team planned to evaluate Delos at the Massachusetts Systems Integration Lab. The objective was to test how well the system could maintain anonymous continuity and contextualize activity across camera views, then use the findings to guide further development.

This paragraph reflects a company announcement. Public descriptions from DHS provide broader program context, but they do not independently validate every product-performance statement in the original release.

How Delos approaches video intelligence

Delos is designed as an AI backend for existing camera infrastructure. Instead of treating every frame as an isolated detection, it maintains a changing operational view of people, objects, movement, and interactions. Lauretta’s privacy-oriented approach avoids facial recognition and focuses on temporary appearance and behavior context rather than biometric identity.

A 2024 DHS S&T market survey on crowd-counting technologies separately describes Lauretta’s use of existing CCTV, a top-down map view, and near-real-time information about crowd count, density, movement, velocity, and flow. This is useful public evidence for the type of operational context Lauretta builds, while remaining distinct from the TSA test announcement.

Airport screening as a physical-AI problem

Airport screening combines perception, process understanding, and human decision support. DHS’s passenger self-service screening fact sheet describes video analytics as part of the interfaces, sensors, and software that can provide travelers with prompts and alert officers when intervention may be needed. The goal is not simply to detect a person; it is to understand whether a real-world process is unfolding as expected.

For more context, read about Lauretta’s 2021 DHS self-screening award, the company’s approach to behavioral analytics without facial recognition, or Beyond Guard.

Continue the evidence trail

Read the evidence-led retrospective on airport self-screening and privacy-preserving video intelligence, then see how detection-to-action workflows connect observations with reviewable operational response.