Practising Clear Explanation with AI

Clarity is a skill you can build like a muscle. But in real life, test subjects are scarce: it's hard to ask a colleague ten times "do you get it now?" AI solves this. It can play the patient beginner, detect the jargon you no longer see, generate analogies to test, and check whether your explanation holds up. It is precisely the tool that was missing to fight the curse of knowledge.

Why AI is an ideal partner for this

The core problem of explaining is that you don't see your own blind spots. An AI assistant, however, can simulate a lack of knowledge: ask it to play someone who knows nothing about your field and to flag every word it doesn't understand. In seconds you get what a real audience would rarely give you so frankly: the list of your "sharp taps."

flowchart LR
    U[You explain<br/>your concept] --> IA[The AI plays the beginner<br/>and flags every snag]
    IA --> R[You rephrase<br/>more simply]
    R --> F[The AI tests:<br/>jargon? logical leaps?<br/>missing example?]
    F --> A[You adjust<br/>and start again]
    A --> IA

Five concrete uses

Five training prompts to copy

1. Ruthless jargon detector

"Here is my explanation: '[text]'. You know NOTHING about this field. Underline every word or acronym you don't understand and every spot where you lose the thread. Don't fix it yet: just point out where it breaks down for a true beginner."

2. The beginner who asks the real questions

"Play the role of an intelligent person who has no knowledge of [field]. I'm going to explain [concept]. After each sentence, ask the naive question that comes to mind. Don't pretend to understand. We stop when you can re-explain the concept to me yourself."

3. Analogy generator

"Give me 5 different analogies to explain [concept] to someone who knows [audience's familiar field]. For each, state what it captures well AND where it stops being valid."

4. Rephrasing test

"Here is my explanation: '[text]'. Rephrase it in your own words, as if passing it on to a colleague. I'll see in your rephrasing what got through and what was lost."

5. Audience adapter

"Rewrite this explanation in 3 versions: (a) for a 10-year-old, (b) for a professional from another field, (c) for a busy executive in 30 seconds. For each version, say which anchor point you chose."

Cautions and limits

AI is a simulator, not your real audience. Three guardrails. First, check the facts: an analogy can be appealing but wrong — you remain responsible for accuracy. Second, transpose to a human: a real listener has context, history and emotions the AI lacks; the final test is always a real person. Third, don't aim for the "perfect," frozen explanation, but for flexibility: knowing how to step back down the ladder of abstraction when the other person disengages. Good AI practice should make you more attentive and more agile, not more mechanical.

A two-week training plan

Week 1 — diagnosis: take three concepts from your field, run each through the "jargon detector," and note your reflex words to ban. Week 2 — calibration: for each concept, have the "beginner who asks the real questions" play, then generate three analogies and test the best one on a real human. Each evening, note the naive question that surprised you most: that's often where your blind spot was hiding.

Summary

AI is the ideal tool against the curse of knowledge because it can simulate ignorance on demand. Five uses: detect jargon, play the beginner, generate analogies, test your clarity through rephrasing, and adapt the level to the audience. Provided you check the facts, transpose to a real human, and aim for flexibility rather than a frozen script. A progressive two-week plan turns these prompts into reflexes — and makes you someone people understand the first time.