How to check your assumptions with AI

Your own assumptions can be hard to notice because they often feel obvious. AI can help name them, but it does not know your hidden motives or the whole context. It can organize a situation description and show what has been added without strong confirmation.

This is different from separating facts and interpretations. There you organize the material. Here you check your own first reading: where you have already added meaning, motive, risk or conclusion.

When this helps

When you describe a situation and move quickly toward a conclusion.
When a decision depends on information you have not checked.
When you have a strong first reading of a situation.
When you want questions before defending your interpretation.
When emotion is an important signal but may also narrow the interpretation.

What to ask the model

Ask the model to check your first reading. Do not ask it to assess your personality or intentions.

Read the description below and help me check my assumptions.

Do not evaluate my personality, motives or emotions. Work only with the material I provide.

Divide the analysis into 6 parts:

1. Facts: what can be pointed to in the description.
2. My first reading: what conclusion or meaning I am adding.
3. Assumptions: what has to be true for my reading to be accurate.
4. Weaker points: which assumptions need checking.
5. Missing context: what needs to be known.
6. Questions: what I should ask or verify before forming a conclusion.

At the end, identify which conclusions would be premature and why.

Description:
[paste description]

What to check yourself

Whether the “facts” are really present in the material.
Whether an assumption comes mainly from emotion, speed or previous experience.
Whether the first reading has started to feel like the only possible reading.
Whether the person or source affected by the situation is missing from the material.
Whether the model added context you did not provide.
Whether the result gives you concrete questions, not only a longer list of doubts.

What can go wrong

The model may name an assumption too confidently.
The user may use the assumption list to defend an earlier decision.
The exercise may look like detached analysis even when the material is incomplete.
In high-risk situations, proper people, sources and procedures are needed.

Better way to use the answer

You describe a situation: “The team is not replying, so the project is not important to them.” The model can separate the fact: the team has not replied for two days. The assumption: no reply means low commitment. The missing context: whether the team has other deadlines, whether they received all materials and whether they know what you expect. This does not solve the situation, but it changes the first move: instead of a judgment, there is a question.

This practice connects epistemic vigilance with metacognition. The model helps reveal the structure of your own reading, but it does not replace checking the situation or human responsibility.

Short rule

Do not ask AI what you really think. Ask it to show what in your reasoning is an assumption.

Continue exploring

TopicAI and cognition
A guide to performance, learning, transfer, offloading and cognitive practice.
ArticleDon't Ask Whether AI Makes Us Dumber. Ask What Kind of Thinking We Stop Practicing
A careful distinction between task performance, learning, retention and transfer when AI becomes part of the process.
ConceptCognitive offloading
Moving part of a cognitive task into an external tool or action.

Suggested citation

Mamczur, F. (2026, July 7). How to check your assumptions with AI. Prompted Psyche. https://promptedpsyche.com/practice/how-to-check-your-assumptions-with-ai/