Imagine an illustrator using AI to prepare descriptions for a new print collection. The descriptions are useful, the administrative work is finished and there is time left for sketching. Yet the next drawing session feels unexpectedly flat. The first task went well; the experience of continuing the work deserves a separate look.
This fictional example raises a question worth asking when arranging AI-assisted work: how does the sequence feel, including the part you do afterwards? A useful draft, a sense of control and interest in the next task are different outcomes. They need room in the review alongside speed and accuracy.
Define the kind of motivation you mean
In their 2000 review of intrinsic and extrinsic motivation, Richard Ryan and Edward Deci distinguish enjoyment or interest in an activity itself from acting to obtain a separate outcome. They also explain that instrumental goals can be personally endorsed and pursued willingly. A task can matter to someone even when doing it is not particularly enjoyable.
For an independent creator, preparing product descriptions may be worthwhile because customers need clear information. Sketching may offer a different satisfaction: exploring a line, shape or idea. Completing the descriptions efficiently and wanting to continue drawing should therefore remain separate questions.
This distinction also helps a business owner discuss work without expecting every necessary task to be entertaining. Ask which activities people value doing, which they mainly value completing, and what they want the freed time to make possible. The answers can differ between colleagues doing the same job.
Keep one experiment’s finding in proportion
A Scientific Reports paper published on 29 April 2025 by Suqing Wu and colleagues reported four online experiments. Its largest analysis included 1,624 participants preparing welcome emails and promotional Facebook posts. Researchers varied AI access across two tasks and counterbalanced their order.
In that study, self-reported intrinsic motivation declined across all four conditions. Participants moving from ChatGPT assistance to solo work showed a small additional decline compared with those working solo throughout, although their sense of control rose. Boredom also increased in all four conditions; its change did not differ statistically between those two groups.
These were questionnaire ratings across two consecutive tasks with online participants. They do not establish a permanent loss of motivation or predict what will happen in a particular workplace. The practical implication I draw is to examine the following task when evaluating an assisted workflow.
Look at the whole sequence
Return to the fictional illustrator. Record what the AI-assisted description task involved, what came next and what the person wanted from that next activity. An observation such as “I completed the descriptions, but struggled to become interested in sketching” is more useful than deciding immediately that the entire workflow has succeeded or failed.
Keep other circumstances visible too: deadlines, interruptions, the difficulty of the drawing brief and how much work has already been done. One afternoon cannot isolate why the experience changed. Comparing the same broad sequence on another occasion may reveal something worth discussing without proving a cause.
The following suggestions are work-design ideas to explore, rather than interventions tested by the experiment. Start with a sequence that already matters to your work. Avoid manufacturing an elaborate monitoring routine that becomes another administrative burden.
Choose what the saved time is for
Specify the purpose of the assistance before judging whether it helped. If the aim is to finish routine descriptions and rest, completion may be the main outcome. If the aim is to make space for a drawing session the illustrator values, the review should include whether that session actually happened and felt worthwhile.
The creator might choose to sketch before preparing descriptions, separate the activities into different sessions, or keep the current order. Another possibility is to use AI for a narrow wording check after drafting the descriptions personally. These are different arrangements to try, with different costs in time and involvement.
There is no reason to assume one arrangement will suit everybody. Ask the person doing the work which trade-off they prefer, then inspect the result. A change that feels more engaging but consumes the time needed for another priority may still need adjustment.
Keep meaningful participation available
Ryan and Deci’s review discusses how motivation relates to autonomy and competence. It treats autonomous motivation as compatible with pursuing a goal someone accepts as valuable. This gives a useful starting point for discussing assistance: identify the participation the person actually wants, rather than equating control with doing everything manually.
For the illustrator, that participation could be choosing which print deserves a fuller description and explaining its intended audience. For a colleague, it could be editing a rough paragraph into something clearer. Another person may welcome extensive drafting help and prefer to contribute elsewhere.
When writing itself is important to the creator, MaryChuks’ guide to keeping a sense of authorship with AI offers a related discussion of involvement. Here, extend the review beyond the finished text to the activity that follows it.
Keep the option to disagree with a suggested workflow. A team member saying that an arrangement leaves the interesting work elsewhere is offering information about task design. Ask which part they would like to shape and whether that change can fit the job’s requirements.
Review output and experience together
For a short, informal review of a real work sequence, consider:
- Result: Did the assisted task produce what was needed, including the review and corrections?
- Continuation: What task followed, and did the person begin and complete the intended work?
- Experience: Which parts felt interesting, worthwhile, frustrating or dull?
- Choice: What would the person keep or change about the sequence next time?
These questions are a proposed conversation aid, not a validated psychological scale. Keep any notes proportionate to their purpose and avoid treating a brief reaction as a permanent description of someone’s motivation.
A useful AI workflow can make necessary work easier while leaving room to notice what happens next. Judge the result, hear the person’s experience and revise the arrangement around the work they need and want to do.
Sources
- Wu and colleagues, Human-generative AI collaboration enhances task performance but undermines human’s intrinsic motivation, Scientific Reports, 29 April 2025. Publisher-provided full text.
- Ryan and Deci, Intrinsic and Extrinsic Motivations: Classic Definitions and New Directions, Contemporary Educational Psychology, 2000.
Sources checked on 27 October 2026. The illustrator, work sequence and suggested review questions are illustrative.
One thought on “After AI Help: Check How the Next Task Feels”