The Pentagon requests $30.3 million for a five‑year AI‑driven lie‑detector program called Polygraph+, focusing on machine‑learning scoring and standoff sensing technology.

The Pentagon requested $30.3 million for a five‑year program to develop an AI‑enhanced lie‑detector system.

The Department of Defense budget proposal names the effort “Polygraph+” or “Polygraph Next.” The initiative will fund research on scoring algorithms that apply artificial‑intelligence and machine‑learning techniques to physiological data collected during polygraph examinations.

Another focus is “standoff sensing,” a method that measures biometric signals from a distance, allowing investigators to assess subjects without direct physical contact. Officials say the approach could speed up screenings for security clearances, military personnel and contractors.

Program scope and timeline

Funding would be spread over the next five fiscal years, with annual allocations tied to milestones such as algorithm validation, sensor integration and field trials. The DoD plans to partner with university labs and private‑sector firms that specialize in signal processing and AI research.

Potential benefits and applications

Proponents argue that AI‑driven analysis could reduce human bias in traditional polygraph tests and improve detection accuracy. Standoff sensing may enable rapid assessments at checkpoints, border crossings and high‑security installations.

Oversight and criticism

Congressional staffers have raised questions about privacy protections and the scientific validity of AI‑based lie detection. The Pentagon confirmed that any deployment will comply with existing privacy regulations and will be subject to independent review by the Defense Science Board.

Critics also note that past polygraph technology has faced legal challenges and limited admissibility in courts. The department acknowledged that the program will undergo rigorous testing before any operational use.

Next steps

The request will be reviewed by the House and Senate Armed Services Committees during the upcoming appropriations process. If approved, the first prototype is slated for testing in late 2027.