AI Is Speeding Up Intelligence, But Not the System Around It

Compression is Already Happening

In some areas, AI is already reshaping intelligence work in meaningful and measurable ways. Former NGA Director Vice Adm. Trey Whitworth (Ret.) has repeatedly highlighted how AI is revolutionizing GEOINT. Full-motion video analysis that once required extensive manual exploitation is increasingly automated and continuous. Project Maven fundamentally changed the economics of GEOINT warfighter support by applying computer vision to operational imagery workflows. Some AI-generated products are being disseminated to senior policymakers with minimal human involvement.

Even before Anthropic’s game-changing Mythos product, SIGINT and cyber operations similarly benefited from AI. NSA's Human Language Technology program automates speaker identification and translation across more than 90 languages—enabling analysts to triage millions of intercepted communications and focus only on the relevant fraction. Cyber Command and NSA increasingly operate in what Former NSA Director Gen. Paul Nakasone (Ret.) called "persistent engagement"—environments where collection, analysis, decision-making, and cyber effects occur continuously rather than sequentially.

Open-source intelligence has arguably made the strongest strides. During recent conflicts, policymakers leveraged commercially available satellite imagery, social media, and public telemetry data in near real time. CIA's OSIRIS platform uses LLMs to synthesize vast volumes of open-source data, deliver summaries, and support analyst engagement through a chatbot. Former Open Source Enterprise Director Randy Nixon argued that these advances enabled OSINT to become “the INT of first resort”—a model for all-source intelligence collection and analysis.

But Compression Is Uneven

GEOINT, SIGINT, and OSINT lend themselves to AI adoption: they are data-rich, measurable, and in OSINT's case, unclassified. Clandestine tradecraft and rigorous analytic tradecraft are harder to accelerate.

The Beginning of a Strategy

Deputy Director Michael Ellis recently said that CIA expects AI to become an everyday “co-worker” for analysts within the next few years. He described a future where AI systems help analysts draft reports, identify patterns across massive datasets, test conclusions, and surface threats. Ellis also said that analysts are already experimenting with how to evaluate, validate, and cite AI-enabled insights. Questions that were largely theoretical only a few years ago are becoming practical tradecraft challenges:

• How should AI-assisted analysis be sourced and validated?
• What level of confidence should accompany machine-generated insights?
• How should analysts distinguish between AI-enabled synthesis and human judgment?
• What standards should govern the use of AI-generated content in finished intelligence?
  

These are important developments because they signal that intelligence leaders are thinking about how technology adoption requires tradecraft modernization.

Coordination, Validation, and Analytic Workflows

Deploying AI tools for isolated analytic tasks (e.g., search, discovery, drafting) is relatively straightforward. An analyst may now receive machine-generated correlations in seconds yet still wait hours or days for cross-agency coordination, sourcing validation, or product approval. Reimagining those surrounding workflows—how information moves, how trust is established, how products are reviewed, and how analysts interact with machine-generated outputs—is substantially more difficult.

The opportunity is enormous but requires redesigning the processes themselves—while continuing to deliver on policymakers’ daily needs.

Policy Integration and Decision Support

The compression challenge becomes even more visible when intelligence intersects with policymaking.

The traditional model of intelligence dissemination was built around periodic delivery and daily briefing cycles like the President’s Daily Brief. However, policymakers now consume intelligence alongside operational updates, open-source reporting, and social media—and make decisions at the edge, often faster than traditional dissemination cycles allow. To adapt, intelligence agencies will need to provide continuously updated context, machine-assisted forecasting, and dynamic collaboration embedded directly into policymaking workflows.

Consider what this looks like in practice: a combatant commander or ambassador can query an AI-enabled analytic system for a continuously refreshed threat picture, stress-test an assumption against alternative scenarios, and receive a validated assessment—all in the 30 minutes before he walks into a meeting with his foreign counterpart. Elements of this reality exist today. During Operation Epic Fury, the 38-day air campaign against Iran, AI synthesized targeting data across the battlespace in real time to support strikes on roughly 13,000 targets in just over a month—a pace of machine-assisted decision-making with no precedent in U.S. operations.

The challenge for intelligence leaders is driving development and adoption to make it systematic, trusted, and governed. As OSIRIS proved, AI-enabled platforms are already beginning to empower policymakers to interact with intelligence this way. In that environment, intelligence is no longer a product delivered to decision-makers. It is the environment in which they decide.

That does not mean abandoning rigor or replacing strategic analysis with real-time reporting. In fact, the opposite is true. As information velocity increases, the value of trusted analytic judgment, validation, and expert perspectives will increase.

The Emerging Risk: Asynchronous Modernization

The danger is that different parts of the system are modernizing at different speeds. Accelerating functions does not eliminate friction between functions. In some cases, it can increase it.

·Faster collection can overwhelm coordination processes.

·Faster analysis can outpace dissemination workflows.

·Automated insight generation can challenge validation, trust, and decision integration.

The Leadership Challenge

Leading companies discovered that AI could not remain a standalone innovation initiative. As AI began reshaping workflows, governance, and strategy, responsibility migrated from CIOs and innovation teams to CEOs and boards. Intelligence leaders must make a similar pivot. Modernization cannot be outsourced to technologists, innovation offices, or isolated teams. Pockets of ‘AI money’ and lists of ‘AI projects’ are technocratic, not strategic.

To fundamentally modernize how the system operates, IC leaders must be directly involved in reimagining the intelligence cycle, redefining tradecraft expectations, reshaping decision models, and aligning institutional incentives. That means three things.

First, leaders must connect the technology agenda to the mission agenda. The private sector learned that AI transformation fails when it is treated as an IT problem rather than a strategic one. The IC faces the same risk. The goal is decision advantage at the speed of policymakers and warfighters. Keeping that mission orientation at the center of the technology agenda is a leadership responsibility that cannot be delegated.

Second, leaders must own the coherence problem. It is not enough to authorize AI investments and track deployment metrics. Leaders must make architectural choices about which parts of the cycle to accelerate together, how to manage the seams between them, and what governance structures are needed to ensure the system absorbs new capability without creating new failure modes.

Third, leaders must redefine what tradecraft means in an AI-assisted environment. The standards that govern sourcing, analytic confidence, and product integrity were built for human workflows. They need to be deliberately redesigned—not abandoned—for an environment where machine-generated insights are embedded throughout the production chain. Deputy Director Ellis’s four questions are the right starting point.

AI may compress intelligence production, but only leadership can compress the distance between insight and decision.

The Cipher Brief is committed to publishing a range of perspectives on national security issues submitted by deeply experienced national security professionals. Opinions expressed are those of the author and do not represent the views or opinions of The Cipher Brief.

Have a perspective to share based on your experience in the national security field? Send it to Editor@thecipherbrief.com for publication consideration.

Read more expert-driven national security insights, perspective and analysis in The Cipher Brief