Overview
- CIA Deputy Director Michael Ellis, speaking Thursday at a Special Competitive Studies Project event, said the agency will roll out classified generative-AI assistants across its analytic platforms in the next few years.
- The push comes as the CIA elevates cyber operations, including turning its Center for Cyber Intelligence into a full mission center and doubling tech-focused foreign reporting to track how rivals like China deploy advanced AI.
- The new assistants will draft key judgments, edit for clarity, check work against tradecraft standards, and triage large data sets to flag patterns for human review, with analysts signing off on the final product.
- The agency has already crossed an internal milestone by producing its first fully AI-generated intelligence report after running about 300 AI pilots last year for tasks like processing big data and translating foreign languages.
- Ellis said the CIA will avoid dependence on any one vendor as the Pentagon’s supply-chain risk label on Anthropic stands after a court declined to pause it this week, reinforcing a strategy to diversify AI suppliers.