When AI Takes the Work You Enjoy but Leaves You Employed

Atlassian reported on November 4 that 82% of surveyed AI users said the technology handled at least some work they enjoyed. A separate 82% said it handled work they disliked. These are findings from the company's own Teamwork Lab survey. [1]

I think this raises a career question that layoff counts cannot answer: what remains of a profession when employment continues but the satisfying acts of doing it disappear? A surviving position can still become a worse job. That deserves attention before someone reaches the point of wanting to leave.

What the delegation figures reveal

The survey ran September 28–October 1 among 1,013 full-time U.S. knowledge workers; the AI questions covered 949 users. Its November publication describes earlier perceptions, rather than a sudden change this week. [1]

Those percentages do not establish equal task volumes or show that respondents regretted the delegation. [1] Someone might gladly surrender an interesting assignment to make room for something better.

But 10% said AI handled all the work they enjoyed. [1] I would take that seriously without pretending it describes every workplace. My concern is what a role offers after the handover: a more rewarding contribution, or a thinner working life.

Meaning need not come from employment. I respect choosing a job for its income and keeping personal fulfillment elsewhere. But enjoying the making, investigating or explaining was never an unreasonable thing to want from a career.

A polished result can still feel less yours

An older experiment described by Penn State on June 5 helps explain why this is more complicated than whether people like AI. Approximately 270 professionals participated in a writing experiment lasting 20–30 minutes. The first task compared manual writing, collaboration with AI and copying AI output; everyone then wrote manually. [2]

According to that account, passive use produced higher initial enjoyment and satisfaction with the result, alongside lower reported ownership and meaningfulness. Collaborative use produced ownership, meaning and confidence ratings similar to manual writing. [2]

That combination matters to my argument. You could welcome an easier task and prefer the finished document while feeling that less of it belongs to your effort. Enjoyment of the process, satisfaction with the output and a sense of contribution are different judgments.

The study measured self-reports, including confidence in writing without AI. It did not demonstrate a loss of writing ability. Its brief experiment cannot establish permanent alienation or predict a participant's career. [2]

I would resist both overreactions: dismissing these feelings because the output looks good, and declaring that AI inevitably destroys meaningful work. The collaborative condition gives the second claim a particularly obvious problem.

The moments a job description misses

By professional authorship, I mean having a recognizable part in how work turns out. That might be the explanation an analyst develops, the direction a designer chooses or the argument a communications worker makes. It need not mean producing every component unaided.

Imagine an analyst who once investigated an ambiguous result and now receives a generated conclusion to approve. The title and deliverable could remain identical. Yet the opportunity to experience the discovery could shrink. This is a hypothetical consequence of redesign, not an outcome measured by Atlassian.

Those small accomplishments matter in my view. A document can be saved, credited and circulated while the person assigned to it has little experience of having made it. The lost moment leaves no vacant position to count.

Equally, evaluating an explanation can require substantial thought. I would not assume that reviewing AI output is inherently inferior work. The distinction I care about is whether the reviewer can question the premises, change the direction and shape the result. A nominal approval step offers a poor substitute if meaningful decisions have already been closed.

Plan for the work you want to continue doing

For career planning, I would treat a major change in daily tasks as a change in the occupation being offered, even if the employer keeps the old title. Someone hired because they wanted to write may discover that the new role chiefly involves selecting and checking generated prose.

That person may enjoy becoming an editor. They may also decide the new work no longer fits. Neither reaction establishes a failure to adapt. My advice is to name the difference precisely before choosing a training course or a different role.

If creating an explanation is what matters, seek learning opportunities that still require developing one. If evaluation is becoming appealing, learn the subject deeply enough to challenge a plausible answer. These are different career directions; treating both as simply becoming better at AI hides the choice.

I would also be wary of a compromise that preserves personally satisfying work only as an unpaid evening activity. Someone can choose a hobby freely. I do not think it is reasonable to call a job redesign an improvement for that worker while assuming they will recreate everything they valued outside working hours.

There is room for new accomplishments

I see room for a better outcome in the collaborative condition. AI could help someone attempt an explanation or project they could not have completed alone, with their choices still shaping the result. That is a possibility, not a lasting career outcome established by this experiment. [2]

The uglier possibility is a job that retains responsibility while shedding participation. I would call that a deterioration, even if output increases and the position survives.

I want progress to include the chance to understand something difficult, make a consequential choice or create something useful. A faster finished product cannot give someone back an experience of discovery they never had. I think that cost belongs in the decision about what to automate.