When Intelligence Judgments Become the Problem

The CIA’s report retractions risk teaching analysts which conclusions are acceptable.

The Central Intelligence Agency says its decision last Friday to retract or revise 19 intelligence reports is a corrective action to remove bias and restore analytic objectivity. CIA Director John Ratcliffe ordered the move after the President’s Intelligence Advisory Board — a panel drawn from outside of the U.S. intelligence community — reviewed hundreds of CIA products and flagged the assessments for withdrawal or “substantive revision”; the CIA posted redacted versions of three of these reports.

Several former intelligence officials who reviewed the posted reports told the media that they questioned the agency’s basis for its actions. They argue the pieces reflected the policy priorities at the time, not analytic misconduct, and they contained no fundamental flaws in tradecraft.

(brightstars/iStock)

I’ve reviewed the same reports. My former colleagues are correct.

Before entering academia, I served as a senior intelligence officer and later as an executive reviewer of the President’s Daily Brief, the intelligence community’s flagship analytic publication. Today I teach strategic intelligence at Georgia Tech, where I have students take on the role of an intelligence analyst and produce finished intelligence deliverables modeled on the real thing. They learn quickly that the challenge is not the template, it’s the discipline: crisp judgments, defensible sourcing, a clean separation of evidence from inference, and the ability to stay out of advocacy when the topic is consequential and contested.

That is the standard I am applying here, and, by that standard, the agency hasn’t made its case.

Some partisan outlets are describing the reports as “DEI-infused intelligence” that sought to advance progressive agendas; the posted reports dealt with topics such as gender roles in extremism, LGBTQ issues in the Middle East, and pandemic-era contraception shortages. As a result, some are seeking to push the story into the broader culture-war narrative.

But the reports do not read like ideological documents. They include clear judgments, cite reporting, and draw implications without telling anyone what policy to adopt. That’s the boundary the intelligence profession is supposed to keep.

And that brings us to the most basic point that is often lost outside the government: Finished intelligence is not a mandate. It is not an administration’s policy. It is not a set of marching orders. It is an assessment — an attempt to bound uncertainty and reduce surprise by telling decision-makers what the evidence suggests and what risks may follow.

Policymakers can dismiss analytic judgments. When I was an analyst, some of my products landed poorly with decision-makers — an experience so common it barely warrants explanation. That kind of friction doesn’t discredit the intelligence; it illustrates the boundary between analysis and policy. One informs. The other decides.

Some assume that if intelligence touches politically sensitive subjects — domestic extremism, social movements, or political rhetoric — then bias must be involved. That is not how intelligence analysis works. Analysts follow questions that policymakers ask and developments that appear relevant to national security. The topics themselves are not evidence of politicization.

The intelligence community has been through a version of this argument before. Last July, Ratcliffe declassified a tradecraft review of the 2016 Intelligence Community Assessment on Russian election interference, describing it as proof that senior officials had “manipulated” intelligence to target Donald Trump. But the document itself did not say that. It described an accelerated, leadership-heavy process and debated judgments under imperfect sourcing, but it affirmed strong adherence to tradecraft and did not repudiate the underlying finding of Russian interference.

The gap between Ratcliffe’s rhetoric and the review’s actual conclusions was the point: A process critique was turned into a political weapon. Friday’s action appears to be repeating that dynamic.

Trump has repeatedly shown little patience for intelligence judgments that contradict his views — on Russia, on Iran, and on other national security questions. Under this administration, the pressure is not conveyed only through public disputes over particular products; it is reinforced by a broader pattern of actions that treat unwelcome analysis as a management problem — reassignments, intensified scrutiny, and the removal of senior analytic leaders.

Over time, that environment threatens — and is likely designed — to change behavior: Analysts hedge, narrow what they will assess with confidence, and steer away from judgments likely to trigger blowback. None of that improves intelligence. It just reduces candor.

Defenders of the CIA’s decision may argue that enforcing analytic standards strengthens the institution. In principle, that is true. Internal reviews and post-mortems are routine parts of the intelligence profession. But those processes occur inside the system, not as public events designed to settle political arguments.

If the goal is to improve intelligence, the conversation should start there. Otherwise, the lesson analysts will absorb is not about tradecraft at all. It will be about what kinds of judgments bring punishment — and that is how candor erodes long before anyone can measure the damage.

Brian O’Neill, a retired senior executive from the CIA and National Counterterrorism Center, is an instructor on strategic intelligence at Georgia Tech. His Safehouse Briefing Substack looks at what’s ahead in global security, geopolitics, and national strategy.

---