DCSA Denies Using Generative AI in Polygraph Modernization

The Defense Counterintelligence and Security Agency announced that its Modernizing Polygraph effort does not currently incorporate generative artificial intelligence, large language models, or AI-based analysis of facial expressions and vocal patterns. This clarification follows reports and declassified slides showing that the Department of Defense, the Air Force, and academic labs have developed AI tools since 2019 to detect deceptive speech and estimate trustworthiness for security clearance vetting.

The project, known as Polygraph+ or Credibility Assessment Modernization, is budgeted at 31 million dollars and utilizes thermal imaging and scoring algorithms. While the Pentagon previously described AI as a force multiplier for human investigators, the agency now maintains that several of these technologies are not part of the active research and development phase.

National security lawyers, scientists, and AI auditors have criticized the initiative, arguing that AI-aided deception detection lacks scientific backing and may violate civil liberties. Critics warn that the system could lead to automation bias and discrimination against individuals with disabilities, potentially derailing military careers based on unreliable AI opinions.