AI Brainstorming: Make Room for a Return to the Problem

The assistant has suggested another page of workshop ideas. You have rejected some, combined two and asked for ten more. Before continuing, it is reasonable to ask what the next batch is meant to resolve. Perhaps you need more options. Perhaps the useful next step is to leave a clearly stated problem alone for a while and return with your own judgement.

Research on creative incubation gives that second possibility a specific context. It concerns what happens after people first encounter a problem and then spend an interval away from directly working on it. It does not establish that every distraction is productive or that an AI conversation becomes better simply because it is interrupted.

What one incubation study examined

Benjamin Baird and colleagues’ Inspired by Distraction: Mind Wandering Facilitates Creative Incubation appeared online on 31 August 2012 and in the October 2012 issue of Psychological Science. The researchers used the Unusual Uses Task, in which participants generated unusual uses for familiar objects.

After an initial attempt, an undemanding task during the incubation interval produced greater improvement on previously encountered problems than a demanding task, rest or no break. The helpful condition was associated with more reported mind wandering, rather than more deliberate thought about the creativity task.

This was a laboratory finding with a defined task. It did not test AI brainstorming, commercial workshops or the quality of finished products. The association with mind wandering also should not be presented as an independent demonstration that any chosen daydream causes better ideas.

Begin with a problem worth returning to

Consider a fictional creator called Ada designing a short drawing workshop. She wants participants to practise noticing shape without feeling that they must produce a polished picture. Her unresolved question is which opening activity communicates that intention.

Ada has already sketched one exercise and reviewed several assistant suggestions. The latest list adds materials and themes, but it does not answer her question about the opening. For this proposed workflow, she writes a short note: find an activity that makes noticing shape the purpose, with no requirement to finish a recognisable drawing.

That note matters to the workflow because it names the question she intends to return to. It does not turn her planning session into a replication of the experiment. A workshop design problem differs from producing unusual uses for an object, and her eventual decision needs its own evaluation.

Before stepping away, Ada keeps the current ideas and her reasons for rejecting them. She can then distinguish a new proposal from an old one she has forgotten. This is an editorial record-keeping suggestion, not a claim that a particular note format produces incubation benefits.

Make the pause a choice rather than a promise

Ada decides to spend a brief interval arranging blank paper for the workshop instead of requesting another list. She leaves the chat closed during that interval. The activity is simply her chosen way to pause this session; the article does not prescribe a universal break duration or claim that arranging paper improves creativity.

If the pause produces no new idea, she has not failed a psychological test. She can return to the original options, revise the brief or seek a colleague’s perspective. There is no need to explain an ordinary planning outcome as evidence of a hidden personal trait.

Nor does the research require her to stay mentally occupied with the workshop throughout the interval. The point of the proposed pause is to stop demanding another answer from the same conversation for a time. Avoid making a new performance target out of whether her mind wandered enough.

Return before reopening the suggestions

When Ada returns, she writes any ideas she currently has before reading the assistant’s earlier list. This sequencing is a proposed way to preserve a separate record of her contribution. It has not been tested here as a method for improving originality.

Suppose she proposes asking participants to draw only the empty spaces between objects. She then compares that proposal with the existing ideas. The relevant question is whether the activity communicates the workshop’s intention and can be explained clearly to the intended participants.

The idea is fictional, and calling it new within Ada’s session makes no claim that nobody has used such an exercise before. A private brainstorming record cannot establish worldwide originality. If she intends to describe a method as uniquely hers, that would require a different investigation.

Ada may ask an assistant to help draft a concise instruction or identify what the instruction assumes about available materials. She should keep her own brief and the source of each proposal visible. An assistant’s favourable reaction is not a recorded response from a workshop participant.

Evaluate the idea after generating it

For the fictional workshop, Ada could review the candidate activity against a few concrete requirements. These checks concern its practical fit, rather than the number of suggestions produced during the session.

• Can she explain the activity without suggesting that a polished picture is required?
• Do the instructions make the observation task clear?
• Are the necessary objects and materials available?
• Does the activity fit the time she has actually allocated?
• What would she ask a trial participant to do or explain before approving it?

A trial would add observations that the brainstorming conversation lacks. Ada would need to record what happened, including confusing instructions or practical difficulties. The pause itself supplies no evidence that the resulting activity works.

There is a related question in MaryChuks.com’s article on design fixation: are the alternatives meaningfully different, or are they variations of one direction? For Ada, that comparison belongs after she has identified candidates. It should not be replaced by a general instruction to keep generating.

Keep the role of the assistant explicit

In this proposed arrangement, the assistant helps formulate and compare options. Ada decides what the workshop should achieve, chooses when to pause and evaluates what to try. That division is an editorial choice for the example, not an empirical account of how every person should collaborate with AI.

Some sessions will need facts checked, a constraint clarified or a specialist consulted more than they need a break. Name the unresolved question first. A pause cannot supply missing evidence, and additional suggestions cannot settle a requirement that nobody has defined.

The incubation study offers a reason to investigate the space between initial work and a later attempt. For everyday AI-assisted creativity, treat a pause as an option worth describing precisely. Keep the problem, record what you return with and judge the eventual proposal against its purpose. The next useful contribution may be a better question rather than a longer list.

Sources

Primary research checked on 31 October 2026. Ada, the workshop and the workflow are hypothetical. No AI brainstorming session, creative exercise or intervention was tested for this article.

• Baird et al.: Inspired by Distraction: Mind Wandering Facilitates Creative Incubation — online 31 August 2012; Psychological Science, October 2012. Author-hosted paper at UC Santa Barbara.