Reviewing Only What Is Written Will Not Reduce Errors── Unless you question the information gap between source documents and materials, review keeps its blind spots

01The operator told an AI that reviewing visible text alone will not reduce errors

In September 2026, the operator handed an AI a single unease. Errors in promotional material review kept recurring. The review process was in place. Yet the same kinds of findings would not go away. The operator's reading was that reviewers were looking only at the information written in the material.

The operator passed a hypothesis to the AI: unless you suspect the difference in information volume between the source document and the material, incidents will follow. The operator also asked the AI to connect this structure to cognitive tendencies and accident case studies from other industries, and shape it into an article.

The article skeleton the AI returned contained concepts from cognitive psychology and accident reports from aviation and healthcare. A workplace unease had turned into a reproducible structural explanation.

02Closing judgment on visible information is a cognitive structure, not a workplace habit

In his 2011 book, Daniel Kahneman reported a tendency that appears repeatedly in human judgment: people build a coherent story from whatever information is at hand and fail to notice that information is missing. Kahneman called this "what you see is all there is."

Applied to material review, the act of opening a material and checking each expression one by one sits squarely on top of this tendency. When the text reads well, the motivation to open the source document and check whether information was left out rarely arises. A well-formed text, by itself, suppresses the signal that something may be missing.

03Overlooking absent information costs more than catching visible errors

An error in a written expression can be corrected once pointed out. A gap in information goes unnoticed, the material is finalised, and it reaches the reader. The reader makes decisions based on what the material contains. If a precaution present in the source document was dropped from the material, the reader acts without knowing that precaution existed.

A WHO report on patient safety estimated that diagnostic errors occur in 5 to 20 percent of encounters between physicians and patients. One of the contributing factors the report cites is a break in the transmission of information. The structure in which information is lost as it moves from source document to material is the same structure: a break in transmission.

In the pharmaceutical domain, regulators have repeatedly noted cases in which safety information was not adequately reflected in materials. The root of the problem was not that information was deliberately cut, but that no one noticed it had been cut.

04The information gap is the total volume of facts in the source that did not reach the material

A definition is needed. The source documents — approval dossiers, package inserts, review reports, published clinical trials — contain a set of facts. The material contains a subset. The difference, the facts that were in the source but not carried into the material, is what this article calls the information gap.

The gap will never be zero. The material has a finite page count, and transferring every fact from the source is neither necessary nor useful. The problem is not that the gap exists, but that its contents go unexamined.

05In material review, the costliest blind spot is missing safety information

In practice, the information gap matters most for safety-related data. Efficacy data is the subject of the material and gets included. But the frequency of adverse reactions observed in a trial, precautions for specific patient groups, and conditions for contraindicated combinations are more likely to be thinner in the material than in the source.

The Institute of Medicine's 2000 report, "To Err Is Human," found that many errors in healthcare were caused not by individual carelessness but by system design. The report estimated that up to 98,000 people died each year from medical errors. Its central message was that the problem is not bad people but good people working in bad systems.

Read in the context of material review, this finding translates directly. The reason a reviewer does not check the source is not negligence; it is a process designed around the material alone. If cross-referencing the source is not part of the procedure, even the most careful reviewer cannot notice the gap.

06Three mechanisms stack up to let the gap go unexamined

Why does the information gap persist? Three mechanisms stack up.

The first is the cognitive tendency described earlier. When the information at hand makes a coherent story, people do not go looking for what is missing. A well-written material suppresses the urge to check the source.

The second is the separation of roles. Reading the source is the job of the person who creates the material; the reviewer sees the finished product. Because this separation looks reasonable, it discourages the reviewer from going back to the source. An unspoken boundary says, "checking the source is the author's role."

The third is the structure James Reason described in his analysis of accidents in aviation and healthcare. Reason argued that an accident does not result from a single failure. It results when holes in successive layers of an organisation happen to line up. In material review, an omission during drafting, a blind spot during review, and a perfunctory sign-off by a manager are holes in three layers. When they align, a material with missing safety information goes out.

07From tomorrow, add the source-to-material gap to the scope of review

Three practical steps.

1. Build a list of safety items from the source first. Before starting the review, extract safety-related items from the source: adverse reaction frequencies, precautions for specific groups, contraindication conditions. Ask an AI to do this and the list comes back in minutes.
2. Compare the list with the material and write out the differences. Identify which items on the list are not reflected in the material. Not every item needs to appear. But to decide not to include one, you first need to see it.
3. Record the reason for leaving an item out. If you decide not to include an item from the gap, write the reason in one sentence. When another reviewer looks at the same material, the rationale is traceable.