Pentagon walks back AI polygraph trustworthiness analysis after criticism
hoofnagle · x · 2026-10-06
According to C4ISRNET, the Pentagon had been preparing AI tools that analyze text, voice and facial patterns to estimate trustworthiness, per declassified slides and prior statements as of July. After press inquiries and criticism, officials now say that AI is not part of the current polygraph revamp.
- The Defense Counterintelligence and Security Agency (DCSA) stated the "Modernizing Polygraph" effort's active R&D "does not incorporate generative AI" or LLMs, and involves no "facial action coding, micro-expression analysis or vocal analyses."
- Critics argue AI mind-reading via speech stress and facial expressions lacks scientific backing, compounds errors, and may violate civil liberties in a domain where few such rights already exist.
More from Safety
- Does OpenAI's Watermarking Degrade Model Quality? Reddit Debates the KLD Cost — ResearchCrafty1804 · 2026-10-06
- OpenAI to apologize again to Australia over Medicare hack as artists warn of copyright 'roadkill' — nordicinst · 2026-10-06
- Researcher Tom Davidson Defends OpenAI's Open-Access Stance Against Anthropic-Style Lockdown — AdrienLE · 2026-10-06
- Has California's SB53 AI Whistleblower Protection Ever Actually Been Used? — tszzl · 2026-10-06
- Docker engineer: agents should hold zero credentials in sandboxes, with MCP gateways as the control point — AI Engineer · 2026-10-06
- Coefficient Giving plans $1B for AI safety in 2026, launches Project Tailwind — tallinzen · 2026-10-06