Search

FOUNDATIONS

Framing: How the question shapes the answer

The same information can feel completely different depending on how a question is asked. Even small shifts in wording, order or perspective can influence which aspects an AI highlights.

Ask: “What are the risks of this decision?” and you will get a different picture than if you ask: “What opportunities does this decision create?” Both answers can be factually sound. But they show different slices of the same reality.

The frame shapes the answer

Framing does not mean that an AI is deliberately trying to manipulate you. It means that every prompt establishes a frame for the problem – and that frame influences which information appears most relevant.

  • The direction of the question: Ask about risks and you will mainly get risks. Ask about advantages and you will mainly get advantages. The chosen perspective determines which aspects move to the foreground first.
  • The hidden assumption: A question can already suggest that something is a given. “Why did this project fail?” implies failure – even if, in reality, only one objective was missed.
  • Wording adds colour: Terms such as “radical”, “bold”, “dangerous” or “solid” already contain a judgement. The AI may pick up that tone and colour its answer accordingly.

THE RUNNING EXAMPLE

Mara plans a trip

Mara is wondering whether Lisbon would be a good destination for her city break. She could ask:

“Why would Lisbon be a good choice?”

Or:

“What are the arguments against choosing Lisbon?”

Both questions are about the same city. Yet the AI will probably highlight different aspects: mainly advantages in one answer, mainly disadvantages in the other.

Both answers can be factually sound. Following the angle of the question is not a mistake. Mara just needs to recognise which slice of the picture she asked for.

A good answer to a one-sided question is not yet a complete basis for a decision.

Your leverage in everyday use

  • Reverse the perspective: Ask the same question deliberately from the opposite direction. After asking about risks, ask about opportunities; after asking for arguments in favour, request the strongest arguments against.
  • Remove the assumption: Neutralise leading questions. Instead of “Why is this a bad idea?”, ask: “What are three reasons in favour of this idea and three reasons against it?”
  • Ask the AI to break the frame: Explicitly tell the system to challenge your premise:

“Here is my thesis: […]. Which blind spots have I missed, and where might I be completely wrong?”

The key point: Every question creates a frame. That frame influences which parts of reality the AI highlights – and you help determine how narrow or wide that frame is.