AI-Assisted Reflection: View a Disagreement as an Observer

A collaborator rejects your preferred cover design. You open an AI assistant and begin explaining why the decision is unfair. Before asking for help, you have already supplied a plot, a culprit and a conclusion. The immediate editorial question is whether that account leaves room for anything you have yet to discover.

Try writing the same situation as an observer describing two people with different responsibilities. What does each person know? What remains disputed? What would clarify the next decision? Research on psychological distance offers a reason to explore this change of viewpoint. Applying it through an AI assistant, however, requires its own careful boundaries.

What researchers mean by taking a step back

In a Psychological Science paper first published online on 10 June 2014, Igor Grossmann and Ethan Kross reported three experiments involving 693 participants. People showed wiser reasoning about another person’s close-relationship problems than about their own. The measures concerned recognising knowledge limits, considering other perspectives, compromise and possible change.

In the later experiments, instructions to adopt a distanced perspective improved reasoning about participants’ own problems relative to an immersed perspective. The authors described this as addressing Solomon’s paradox: the gap between reasoning about oneself and about others.

These were specific experimental tasks involving relationship dilemmas. They did not test AI-mediated workplace disagreements, prove that an observer knows the truth or establish that the resulting advice produces better long-term outcomes. The finding invites an experiment with perspective; it does not appoint a machine as an impartial judge.

Clearer reasoning need not feel more comfortable

An earlier paper by Kross and Grossmann, published online on 4 July 2011 and in the February 2012 issue of Journal of Experimental Psychology: General, examined job prospects and anticipated political developments in two experiments. Distanced participants more readily acknowledged uncertainty and the possibility of change.

The study did not find a corresponding effect of distancing on mood. Its authors also left open how people could implement the process outside the laboratory. Feeling better and reasoning differently should therefore remain separate questions. Neither this paper nor the 2014 study validates the AI exercise proposed below.

Keep the disagreement intact while changing the viewpoint

Consider a fictional small publisher working with an independent illustrator on a poetry collection. The illustrator prefers a detailed cover. The publisher asks for a simpler version because the title appears difficult to read at the intended online display size. The illustrator feels that the character of the work is being flattened.

An initial account might say that the publisher does not respect the illustrator’s style. That sentence contains an interpretation of motive. A more useful observer’s account would identify the requested change, the reason actually given and the illustrator’s concern about losing particular visual features. None of those facts requires pretending that the disagreement is trivial.

For this invented project, the illustrator could first write a short account without AI: the publisher requested a simpler cover; the stated concern was title legibility; the illustrator wants to preserve the hand-drawn border. They have not yet agreed which display size to assess or which elements could change.

That version makes a practical gap visible. The parties need a shared example of the intended display. An observer’s perspective should preserve the missing detail rather than fill it with a reassuring explanation about either person’s character.

Give the assistant a limited editing job

If the illustrator chooses to use an assistant, a proposed brief could ask it to rewrite the account in the third person while preserving every supplied fact, marking interpretations and listing unanswered questions. Use fictional names and only the project details needed for the exercise. Read the result against the original notes.

Reject any addition that the publisher is jealous, the illustrator is oversensitive or the audience dislikes detailed art. None follows from the fictional record. A sympathetic sentence is still an unsupported addition if it supplies a private motive or a market result.

The output should leave the illustrator with questions to investigate: What size will readers initially see? Which part of the title becomes unclear? Could the border remain while the lettering changes? These are proposed questions for this example, rather than outcomes obtained in a study or an actual product test.

There is also a stopping point. Once the account and questions are clear, more generated viewpoints cannot establish what the publisher meant. Ask the collaborator, examine the artwork together and update the account with the answer. An imagined response should stay labelled as imagined.

Return to the decision in your own voice

After taking the observer’s view, the illustrator should write a first-person decision. For example: I will prepare a version with clearer lettering while retaining the border, then compare both at the agreed display size. That decision connects a concern, an action and a check without requiring either party to concede a judgement about their personality.

The publisher might still reject the border. The illustrator might decide the revised commission no longer fits the intended work. This exercise does not require agreement or automatic compromise. Its value in this fictional process is making the options and remaining disagreement explicit.

MaryChuks.com’s article on asking for advice about the next version offers a related next step: seek a concrete suggestion tied to a defined creative goal. Here, first clarify whose goal is being discussed and what information supports the proposed change.

Judge the exercise by what becomes answerable

For an ordinary creative disagreement, review the rewritten account with three practical questions:

• Can I point to the observation behind each factual sentence?
• Have I preserved a concern from each person without inventing either person’s motives?
• Does the next action gather information or test a specific option?

These are editorial checks, not a psychological assessment. The useful outcome is an account you can correct and a next conversation you can actually have. A little distance may create room to reconsider; the responsibility for deciding and speaking remains yours.

Sources

Primary sources checked on 29 October 2026. The publishing disagreement and AI-assisted exercise are original fictional applications; neither paper tests AI assistance.

• Grossmann and Kross: Exploring Solomon’s Paradox — Psychological Science; online 10 June 2014, August 2014 issue.
• Kross and Grossmann: Boosting Wisdom — Journal of Experimental Psychology: General; online 4 July 2011, February 2012 issue.