The CIA plans to deploy AI “co-workers” across all analytic platforms within two years. The initiative signals a shift toward embedding generative systems directly into classified intelligence workflows.
Deputy Director Michael Ellis outlined the plan at a Special Competitive Studies Project event in Washington, DC, according to Politico. The tools will assist analysts with routine tasks such as drafting reports and identifying patterns across intelligence datasets. Ellis said human officers will retain authority over critical decisions despite increased automation.
Will AI Co-Workers Reshape Intelligence Analysis Workflows?
The move reflects broader adoption of AI across national security agencies, where speed and data processing capacity are becoming decisive factors. Intelligence operations increasingly rely on real-time analysis of vast datasets, including digital communications and blockchain activity. Compared to earlier systems limited to data retrieval, agentic tools can now execute multi-step reasoning tasks with minimal supervision.
But the rollout comes amid tensions between government agencies and private AI providers. The U.S. administration recently halted use of certain third-party models over security concerns, pushing agencies to develop or control their own systems. This shift suggests a preference for sovereign AI infrastructure in sensitive environments.
“Within the next couple of years, we will have AI co-workers built into all of the agency’s analytic platforms,” Ellis said.
He added that reliance on external vendors could constrain operational capabilities, emphasizing the need for internal control over advanced technologies.
The CIA is also expanding its use of blockchain analytics for counterintelligence purposes, viewing digital assets as part of the strategic competition with China. Ellis noted that the technological gap between the U.S. and China has narrowed significantly in recent years, increasing pressure on agencies to accelerate innovation.
As deployment timelines advance, focus will shift to how effectively AI systems integrate with classified data pipelines and whether internal models can match or exceed private-sector capabilities, with early pilot outcomes emerging as the next catalyst.
