Intelligence: what should the state do with information before it becomes a crisis?
In March 2025, the British government published a public explanation of uncertainty in intelligence assessment. It distinguished the likelihood of an event from confidence in the foundations supporting that judgement. This may sound like a linguistic issue, but it affects state decisions: a danger may be assessed as likely while the supporting information remains limited and susceptible to change. Remove that distinction from a report and the decision-maker receives a certainty the analysis never produced. Official uncertainty guidance.
Iraq has a direct interest in this distinction, because discussion of intelligence often concentrates on possessing information or obtaining it quickly. Information can arrive without being understood; it can be understood without reaching the authority able to act; or it can reach that authority without becoming a decision. An official may also make a sound decision on the available evidence before circumstances change. What should a state do between receiving a signal and the emergence of a crisis? How can this process be held accountable without demanding knowledge of the future or excusing every institutional failure through secrecy?
Information, assessment and decision are different responsibilities
Iraq’s Constitution provides a clear basis for separating these responsibilities. Article 9 assigns the National Intelligence Service the collection of information, assessment of threats and provision of advice to government, under civilian control, legislative oversight, law and human rights principles. Article 84 connects agencies’ work to legally defined duties and powers. Advice is not equated with political decision, nor does secrecy provide an independent mandate. The constitutional text. Intelligence evaluation should therefore begin with the quality of advice enabling government to discharge its responsibilities, rather than turning an agency into a parallel government.
Information is a description or signal requiring examination; assessment is a judgement connecting information with assumptions. A decision selects action amid resources, values and risks that may extend beyond an analyst’s remit. Confusing these levels creates two routes to error: officials can present a political preference as an intelligence finding, or analysts can act as though their judgement obliges government to adopt a single option. Professional independence does not mean separation from the state’s questions. It means retaining the capacity to give an answer different from the one the questioner preferred.
Recipients must therefore understand what has been established, what inferred and what remains unknown. A report need not disclose every piece of raw material, but it should expose the boundaries of its argument sufficiently for review. Several sources tracing back to one account are not independent confirmation merely because the account is repeated. If a central assumption changes, its effect on the conclusion should be stated rather than leaving the report’s headline unchanged. These are standards of reasoning and documentation, not a description of collection methods or a disclosure of security operations.
Iraq’s place in the lessons of British assessment failure
The British Iraq Inquiry, published in 2016, concluded that the presentation of intelligence and assessments concerning Iraq’s capabilities and intentions conveyed certainty that did not reflect the limitations of the evidence. It emphasised the independence of the Joint Intelligence Committee and the distinction between assessing intelligence and constructing arguments for a policy. These are findings from a historical inquiry into British decisions, not an evaluation of Iraqi agencies today. They have particular relevance to Iraq because the consequences of those decisions fell on its territory and society. Executive summary, particularly paragraphs 840–844.
The lesson extends beyond acquiring better information. An institution may produce a report containing qualifications, only for those qualifications to disappear as it becomes political messaging or an executive decision. Core judgements and their limitations must therefore remain connected through successive uses. If an assessment supports a policy, a later independent review should be able to compare what analysts said with what was attributed to them. Preserving this separation also protects decision-makers from building major commitments on certainty that existed only in the presentation.
The more recent British guidance adds a useful distinction between probability and analytical confidence without pretending to eliminate doubt. Analysts may consider an outcome likely while explaining that the judgement rests on fragile foundations; alternatively, they may have strong confidence in evidence indicating that a particular danger is unlikely. My proposed implication for Iraq is a shared institutional language understood by sender and recipient, accompanied by explanations of its limits. This is not a recommendation to copy British terminology verbatim or attach precise numerical probabilities to judgements unsupported by quantitative data.
Warnings that fail to change preparedness
A different review demonstrates that state failure can occur after a risk has been recognised. The first report of the British Covid-19 Inquiry, issued in July 2024, criticised aspects of risk assessment and preparedness, inadequate challenge to prevailing assumptions and groupthink within the system. This was a review of public health and emergency arrangements, not an intelligence agency assessment. Its relevance concerns the relationship between knowledge and preparedness: having experts and reports does not ensure that plans are tested or their limits understood. Resilience and preparedness report, chapter 6.
For Iraq, I propose distinguishing a record of warnings from a record of institutional responses. The first would establish what was said, when and with what confidence; the second would identify who received the assessment, what was decided, why, and when the decision was to be reviewed. Sensitive details need not be published, but should be preserved under authorised oversight. This would help identify whether failure occurred in gathering information, interpreting it, allocating responsibility or implementing a decision. Without that separation, an institution that gave warning may be punished, while one that failed to act can hide behind the claim that the warning was not certain.
Waiting for doubt to disappear is not rational for every risk. Some preparedness measures are inexpensive and reversible, while the cost of surprise may be high. In other situations, premature intervention may cause more harm than the anticipated danger. Choosing between these possibilities is a political and legal responsibility that should benefit from assessment rather than replace it with demands for a definitive phrase. A useful question is therefore: which decision changes if this information is correct, and which additional knowledge would justify the cost of obtaining it? That is a question about resources and accountability, not an invitation to expand agency powers without limits.
A warning also needs an identifiable decision-maker. When an issue spans security, health, energy and finance, regular meetings between representatives are insufficient unless responsibility for leading the response and resolving conflicting priorities is clear. Relevant information may already exist in a civilian institution. Using it lawfully does not require turning that institution into a surveillance arm, but cooperation that preserves its remit and data protections. Mission boundaries matter so that comprehensive security does not become a justification for subordinating every public activity to a single security logic.
How should assessments generate learning?
Performance review must guard against hindsight. After a crisis, it is easy to discover an earlier signal and portray it as an obvious warning while ignoring other signals available at the time. I would therefore assess a sample of judgements against the information available when they were produced, rather than only against what became known later. Reviews should include warnings that did not materialise, unrecognised dangers and judgements revised in response to new evidence. The purpose is calibration and learning, not rewarding someone who warns about every possibility until one proves correct.
Institutions should also permit documented professional disagreement with a prevailing judgement without making disagreement an end in itself. Useful dissent identifies the different evidence or assumption and explains its significance for the decision. A proliferation of unexamined opinions may increase confusion rather than reduce it. Leadership can commission independent reviews of assessments carrying major consequences, while protecting information and clearly assigning responsibility for the final version. These are governance proposals, not descriptions of existing Iraqi arrangements or claims that such measures alone can prevent surprise.
This analysis connects to internal security, intelligence and law enforcement through a specific principle: sound knowledge does not give its holder unlimited authority to act. A good assessment reduces ignorance and reveals its boundaries, leaving decisions to their lawful authority and allowing the relationship between them to be reviewed. The more traceable that relationship becomes, the less judgements of success or failure depend exclusively on institutions’ accounts of themselves.
The value of information before a crisis lies in improving a lawful decision while time for action remains, rather than in its secrecy. A state that preserves professional qualifications, assigns responsibility for responses and reviews its mistakes may still be surprised. It will, however, be better able to learn than one that hears only assessments matching its preferences. The requirement is not a promise of intelligence that knows everything, but a system that prevents available knowledge from being lost between report and decision.
Ali Zuweid — Researcher in strategic affairs and national security
Cite this article
Ali Zuweid. Intelligence: what should the state do with information before it becomes a crisis?. 11 November 2026. https://alizuweid.com/articles/intelligence-assessment-warning-and-state-decisions/?lang=en
