AI Exposes Management Issues More Than Employee Problems

A new Atlassian study has found that honesty about AI use at work may be backfiring, with employees who disclose using AI judged more harshly than peers producing the same quality of work without mentioning it.

The research found that 94% of US knowledge workers now use AI at work. However, in a controlled experiment, workers who disclosed AI use were seen as lazier and were less likely to be recommended for high-visibility work, even when the final output was the same.

In response, Fineas Tatar, leadership expert and co-founder of Viva Talent, says the problem is not that employees are using AI. It is that many companies have not updated how they define good work.

“AI is exposing a management problem more than an employee problem,” says Tatar. “If two people produce the same quality of work, but the person who used AI is treated as less capable, the workplace is still rewarding visible effort over useful outcomes."

“That creates a confusing message for employees. Leaders say they want efficiency, but people may still feel they have to prove they did everything manually in order to be trusted.”

Tatar says many companies are encouraging AI adoption in theory, while still relying on old assumptions about what productivity should look like. That can push employees to use AI quietly, rather than openly, because they worry the tool will be held against them.

He says employers should stop asking, “Did this person use AI?” and start asking, “Was the work accurate, useful, well-judged, and properly reviewed?”

Fineas recommends leaders pressure-test their AI culture by looking at:

• Outcome over optics: Are people judged on the quality and usefulness of the work, or on how manual the process appeared to be?
  

• Disclosure without punishment: Can employees explain where AI helped without being treated as lazy, careless, or less skilled?
  

• Clear human judgement points: Has the company agreed where human review, context, relationship management, and final decisions still need to sit?
  

• Manager consistency: Are leaders encouraging AI use publicly, but penalising it in feedback, project assignments, or promotion decisions?
  

• Workflow value: Is AI being used to remove low-value friction, improve preparation, and speed up follow-through, or is it simply creating more content for teams to manage?

“For fast-growing companies, unclear AI norms quickly turn into inconsistent briefs, duplicated work, quality-control issues, and managers becoming the final checkpoint for everything,” Tatar adds.

“Most teams do not need more vague encouragement to ‘use AI’. They need clearer standards for what good AI-supported work looks like. If leaders do not define that, employees will make their own rules, and some will simply stop being honest about how the work gets done."

“The companies that benefit most from AI will be the ones that treat it as part of better work design: clearer delegation, stronger review habits, better preparation, and less low-value admin. AI should remove friction from the workday, not create a new trust problem between managers and teams."

“The goal is not to make people look busy. The goal is to help them spend more time on judgement, communication, and work that actually needs a human.”