Before you believe it: what is this message actually saying?
Information judgment · Lesson 1: Unpack a claim
Estimated time: 15–20 minutes. No prerequisites.
Teaching note: All organizations, figures and messages below are fictional examples, not research findings.
What you will learn
Given a news item, advertisement, workplace message or AI answer, you will be able to:
- Restate its central claim accurately in one sentence.
- Distinguish checkable factual claims, evaluations, inferences and recommendations.
- Identify unstated assumptions connecting reasons to conclusions.
- Ask specific verification questions and make a provisional judgment proportionate to the evidence.
This lesson focuses on analysis. It does not require statistical testing or determining an author's motives.
1. Start with a message
After introducing AI, a company improved average work efficiency by 40%. People who still do not use AI will soon be left behind.
Before searching or deciding whether this is true, write down:
- What does the message explicitly say?
- What reason does it offer?
- What is the first thing you want to verify?
Keep your answer and revise it after the lesson. This first response is not graded.
2. One message can contain different kinds of statements
A checkable factual claim
“After introducing AI, a company improved average work efficiency by 40%” claims something happened that can be investigated.
A factual claim is a claim that can be checked, not a fact that has already been established. Verification may find it accurate, inaccurate, missing qualifications or unresolved. A precise number does not automatically make it credible.
An evaluation
“This is a satisfying improvement” evaluates a result. By whose standards? Against which target? Were costs included?
Evaluations can have support, but they answer a different question from “How much did the measure change?”
An inference or prediction
“People who still do not use AI will soon be left behind” predicts an outcome. The preceding number may be offered as support, but the connection still needs an argument.
How soon is “soon”? Does “left behind” mean losing a job, earning less or having a task automated? Without a clear meaning, the prediction is difficult to check.
A recommendation
“Therefore, you should buy this AI course immediately” asks the reader to act. Even accurate efficiency data would not establish whether this particular course works, whom it suits, what it costs or what alternatives exist.
The original message does not explicitly recommend buying a course. It may create urgency, but we should not add a purchase recommendation or assume the author is selling something.
Categories can overlap. “Demand will fall next month, so reduce purchasing” combines a prediction with advice. Labels help reveal a statement's role; they are not mutually exclusive boxes.
3. Connect the reason to the conclusion
Reason offered: A company improved average efficiency by 40% after introducing AI.
Conclusion the reader is asked to accept: People who do not use AI will soon be left behind.
Connections that have not yet been established:
- AI chiefly caused the improvement, rather than staffing, processes, task difficulty or measurement changes.
- This company's result applies to the other occupations and people covered by the message.
- The measured efficiency gain is sufficient to cause non-users to lose work or competitiveness.
- That outcome will occur within the period described as “soon.”
These are connections to investigate, not assumptions we have already disproved. Further evidence might support some of them. The message itself has not supplied that evidence.
“After” describes timing. “Because of” proposes causation. One cannot simply substitute for the other.
4. Say what is missing
| What to check | Specific question | Why it matters |
|---|---|---|
| Source | Which company? Where is the original report? Who measured the outcome? | Establish whether the message accurately represents its source |
| Measure | Does efficiency mean output per hour, speed or a subjective rating? Does it include quality and rework? | The same word can describe different outcomes |
| Comparison | What baseline and period produced the 40% figure? | A change needs a comparison to be interpretable |
| Sample | Which people and tasks were included? Were people who dropped out counted? | Results may not apply to everyone |
| Distribution | Did most people improve, or did a few improve a great deal? | An average alone does not describe individual outcomes |
| Explanation | What else changed? Is there a suitable comparison group? | Assess whether AI can reasonably be credited with the change |
| Prediction | What does “left behind” mean? What evidence covers other occupations and longer periods? | One internal company result cannot by itself support a universal prediction |
You do not need to investigate everything at once. Identify the central claim and source, then prioritize questions that matter to your decision. Casual reading may warrant withholding judgment. Quitting a job or paying a large course fee calls for stronger support.
5. A worked provisional judgment
The message claims that average work efficiency rose by 40% after a company introduced AI, but it gives no checkable source, measure definition or sample details. Even if the number is accurate, it does not by itself establish that AI caused the improvement or that everyone who avoids AI will soon be left behind. I would first examine the original material, how the outcome was measured and whom it applies to, then consider its relevance to my work.
This does not say “AI is ineffective” or “the message is definitely false.” It identifies what remains unverified and which conclusions exceed the support provided.
Feeling anxious does not make your judgment wrong. The next lesson examines urgency and emotional language. Here, focus on unpacking the message.
6. Exercise one: identify each statement's role
Another fictional message:
A. A city library says weekend visits increased by 30% this month compared with last month.
B. This shows that residents have developed a reading habit.
C. This is the year's most successful public cultural project.
D. Other cities should immediately copy the approach.
Your task: Label each sentence as a factual claim, inference, evaluation or recommendation, giving a reason. Which sentence most directly requires distinguishing “checkable” from “verified”?
Explore the answer and explanation
Explanation
- A: Factual claim. Check both whether the library said this and whether its records support the figure. The exercise supplies a report of a claim, not its verification.
- B: Inference. More visits do not directly establish a reading habit. Visits are not distinct visitors, and habits require evidence about behavior over time.
- C: Evaluation. “Most successful” needs a comparison set and criteria. With explicit criteria, aspects of the evaluation could also be checked factually.
- D: Recommendation. Copying the approach requires considering local conditions, costs, outcomes and alternatives. A growth figure does not determine what another city should do.
A most directly illustrates the distinction. Attribution to a library does not by itself confirm a number.
Common mistakes: Treating every number as an established fact; assuming recommendations need no evidence; treating every evaluation as an error. Ask what support each statement needs instead.
7. Exercise two: find missing connections
Fictional message:
After adopting a four-day workweek, a team increased its monthly output by 15%. Therefore, every team should adopt a four-day workweek.
Your task: State the reason and conclusion. Identify at least two assumptions to test and one question to prioritize.
Explore the answer and explanation
Explanation
Reason: One team's monthly output rose after a schedule change.
Conclusion: All teams should adopt the schedule.
Assumptions to investigate include whether the schedule caused the increase; whether quality declined or unrecorded overtime rose; whether other teams have sufficiently similar tasks and conditions; and whether overall benefits outweigh costs.
One useful first question is: “Did staffing, task difficulty or output measurement also change?” If your decision concerns adoption, you might first investigate quality, costs or results in other teams. The right priority depends on the decision; there is no single mandatory order.
Common mistake: Naming an alternative explanation and treating it as the proven cause. Alternatives are possibilities to distinguish with evidence.
8. Exercise three: independently review a promotion
Fictional message:
A learning platform says that participants who completed its training program increased their average test scores by 25%. This proves that anyone who enrolls can significantly improve their learning ability. Stop wasting time: enroll today.
Your task: Write an 80–130-word review containing the central claim, the reason offered, at least two gaps and a bounded conclusion. Separately identify new evidence that could change your judgment.
Explore the answer and explanation
Example answer
The platform uses average score changes among program completers to support a claim about everyone who enrolls, then urges immediate action. I would check the original data, whether the tests are comparable, how the 25% was calculated and whether non-completers were excluded. Test scores are also not equivalent to broad learning ability. The message does not provide enough support for its promise to all participants. Before enrolling, I would examine the teaching method, intended audience and complete evaluation rather than rely on this claim alone.
Evidence that could change the judgment: an evaluation covering all enrollees, documenting withdrawals, using comparable tests and including a reasonable comparison group could strengthen or weaken a claim about a specific effect. It would still not automatically show that the program works for “anyone.”
Other answers can be reasonable. “Enroll today” is an explicit call to action, not evidence that the program is ineffective. You can remain interested while rejecting an overbroad promise.
9. Check your reasoning, not just your score
Review your independent answer:
If an item is missing, revise that part. Completing labels or revealing an answer does not demonstrate mastery.
10. A reusable information review
Next lesson: What emotion is this message prompting, and what does it want you to do? We will add emotional language and calls to action to this analysis. Later units address statistical framing, source verification and causal reasoning.
Entries stay in this browser, separately for each language. Export a copy for backup.