AI Daily — September 26, 2026
2026-09-26
anthropic.com
Anthropic Launches Life Sciences Verification Program
Anthropic has introduced a verification program aimed at life sciences organizations, likely to gate access to more capable or less restricted Claude capabilities for vetted researchers and institutions. This follows a broader industry trend of tiered access controls as frontier models become increasingly capable in biology and chemistry domains. The program signals Anthropic's continued focus on responsible deployment in high-risk scientific fields.
technologyreview.com
Pentagon Seeks $30M for AI-Powered Polygraph Program
The Department of Defense has submitted a $30.3 million budget request over five years for 'Polygraph+', a program to replace traditional polygraph scoring with AI/ML-based algorithms and add 'standoff sensing' — physiological signal capture without physical contact. The technical credibility of AI-enhanced lie detection remains contested in the scientific literature, making this a significant policy and ethics flashpoint. The scale of the investment and the standoff sensing component suggest ambitions well beyond incremental improvement to existing systems.