Physical AI creates value when an observation helps shape what happens next. That requires a connected chain from perception to interpretation, communication, and accountable action. Public DHS documents provide two concrete examples: notifications generated from existing CCTV and passenger guidance within self-service airport screening.

Detection opens the action loop

The DHS crowd-analysis market survey describes Lauretta using existing CCTV to detect and notify security personnel of ongoing threats. It also describes live support and insights delivered through a dashboard, API, Telegram alerts, and prediction systems. These outputs are the bridge between an analytic signal and the people or software responsible for an operational response.

Passenger guidance is an action loop

A separate DHS fact sheet describes video analytics tools for passenger self-service screening. The tools provide the interface, sensors, and software that communicate what a traveler needs to do. Individual stations prompt passengers through the steps and alert them if they forgot to screen something, turning observation into immediate guidance rather than a report reviewed later.

Human expertise strengthens the system

The same DHS fact sheet says Transportation Security Officers maintain oversight and can assist passengers who cannot resolve an issue. Automation can make routine instructions more consistent and give people more focused opportunities to intervene, with responsibility, escalation paths, and human judgment built into the workflow.

A practical perception-to-action architecture

The public record supports four connected stages: observe an environment, interpret a relevant condition, deliver a prompt or alert through an operational interface, and route unresolved situations to a responsible person. Dashboards support review, APIs connect other systems, direct alerts support timely attention, and passenger prompts support self-correction at the point of activity.

An evidence trail that accelerates learning

The DHS crowd-analysis survey describes post-event analysis as well as live notifications. When a workflow preserves reviewable context—what the system observed, which condition triggered an alert, what information reached an operator, and what action followed—each event can improve future understanding. Proportionate retention, access controls, and human interpretation make that evidence useful and trustworthy.

Where evidence can unlock the next stage

The documents establish concrete capabilities, prototypes, and workflow concepts. Public evaluation of accuracy, staffing effects, autonomous planning, closed-loop control, production outcomes, and comparative performance can extend that foundation. Each measured result clarifies which actions the system can support next and under what operating conditions.

How teams can advance the complete loop

Teams can test the complete loop: which conditions are recognized, how quickly and reliably information reaches the right person, whether the interface provides enough context, how uncertain signals are resolved, and when escalation adds value. That end-to-end evidence turns a promising analytic feature into an operational system that can learn, improve, and earn a broader role over time.

Related reading

Lauretta and TSA video-analytics collaboration

Lauretta’s DHS airport self-screening award

Beyond Guard

Sources

Crowd Analysis Technologies Market Survey Report — U.S. Department of Homeland Security, 2024-06-24

Passenger Self-Service Screening — U.S. Department of Homeland Security, Science and Technology Directorate, 2024-03-04

A Self-Service Screening Option is Coming to an Airport Near You — U.S. Department of Homeland Security, Science and Technology Directorate, 2023-11-30