AIYING LIVE · V6.0
What If AI Had to Wait for Your First Idea?
See differently. Discover possibilities. Make them real.
By Ravindra Dastikop
You are planning a workshop.
Before writing anything yourself, you ask AI:
“Suggest ten engaging activities.”
Within seconds, you have a polished list.
You select three, improve the wording, and begin preparing.
But one question remains:
What might you have noticed before that list gave your thinking a direction?
A useful capability—and a tension
AI-generated ideas can help people create better work. Yet individual improvement and collective variety do not always move together.
In a 2024 short-story experiment, researchers Anil Doshi and Oliver Hauser found that access to generative AI ideas improved evaluations of individual stories, particularly for less creative participants. However, the AI-assisted stories became more similar to one another. This finding concerns a particular writing experiment; it does not establish that every use of AI reduces originality. [1]
It raises a practical question:
Could we use AI to expand our starting points while preserving the differences we bring?
The hidden assumption
To benefit from AI, invite it into the task as early as possible.
That seems reasonable. Early assistance can save time.
But it also lets the assistant supply the first examples, categories, and possibilities you encounter.
Let us challenge the sequence.
PO: AI must wait for your first idea
Your contribution need not be impressive. It might be:
- Something you saw a learner struggle with.
- An awkward moment in a client meeting.
- A solution you tried that failed.
- A question you cannot yet answer.
These are starting materials that a generic request does not contain.
Move from “generate for me” to “explore from here”
Imagine this fictional workshop-planning situation.
Instead of requesting ten activities, you first write:
“In my previous workshop, participants copied the demonstration successfully but struggled when one detail changed. I want an activity that reveals whether they can adapt.”
Now AI has a particular difficulty to explore.
You could ask:
“Use this observation to propose three activities that test adaptation. For each, explain what the participant must decide independently and what observable result would demonstrate understanding.”
The difference is consequential.
“Engaging activities” asks for a broad category.
“Can they adapt when one detail changes?” establishes a specific purpose.
Your observation gives the generation somewhere useful to go.
Try the three-minute first contribution
Before your next AI conversation, spend three minutes completing these sentences:
- I noticed…
- What makes this difficult is…
- My first possibility is…
Then use this prompt:
Here is my observation, difficulty, and first possibility.
[Paste your three sentences here.]
Before suggesting solutions:Treat my observation as context. Treat my explanation and first idea as claims we can question.
- Identify the assumption in my first possibility.
- Ask one question that could materially change our direction.
- After I answer, generate three substantially different approaches.
- Include one approach that challenges my starting idea.
- Help me choose a small test using criteria we state explicitly.
That last instruction matters. Your first contribution should provide direction without becoming an unquestionable commitment.
Test the sequence
Use a modest task, such as designing a learning activity. Try the two routes in separate fresh conversations, using the same task and assistant where possible.
| Route | Starting point | What to examine |
|---|---|---|
| A | Ask AI for ideas immediately | Which suggestions address your actual difficulty? |
| B | Write your three-minute contribution first | Does the output use your observation meaningfully? |
Compare the results using three questions:
- Which proposal addresses a specific need?
- Which introduces a useful direction you had not considered?
- Which can you test with a clear sign of success?
This is a personal experiment, not proof that one sequence always works better. Route B also supplies more context, so the comparison does not isolate timing alone.
Where could the situation generate its own next possibilities?
Your first observation can become the beginning of a continuing inquiry.
Suppose the activity reveals that participants can adapt one step but cannot explain why it works. That result creates the next question:
“What activity would help them explain the principle behind the change?”
The sequence becomes:
Notice → question → generate → test → notice again.
This is a proposed way to organise the work. Its value depends on what your actual tests reveal.
Accomplish what matters
The accomplishment is an idea developed from something you noticed, improved through questioning, and checked through action.
For your next task:
- Write one observation before opening the AI conversation.
- Use AI to question and expand it.
- Choose one possibility to test.
- Record what happened and what it suggests next.
Source and evidence note
[1] Doshi, A. R., and Hauser, O. P. (2024). Generative AI enhances individual creativity but reduces the collective diversity of novel content. Science Advances, 10(28), eadn5290.
The study provides the research context. The PO provocation, fictional workshop example, and three-minute exercise are proposed applications developed for this post. They are not outcomes established by that study.


No comments:
Post a Comment