Science

AI time savings may be fueling workplace guilt

The Conversation argues that discomfort over AI-assisted work reflects workplace cultures that still equate effort with value.

Priya Raghavan

By Priya Raghavan · Science Reporter

3 min read

AI time savings may be fueling workplace guilt
Photo: Phys.org

AI tools can cut hours from routine office tasks, but some workers may feel uneasy rather than relieved when the work gets done quickly. Paul Jones, writing for The Conversation, argues that this discomfort reflects older beliefs about work, effort and personal worth more than the technology itself.

Jones describes the feeling as “productivity guilt”: the sense that time saved by technology must be justified, filled or repaid. In his account, AI exposes a tension already present in many workplaces, where busyness, long hours and fast replies are often treated as signs of commitment and competence.

AI systems can draft emails, summarize reports, organize ideas and assist with tasks that previously took much longer, according to The Conversation. That can improve efficiency, but Jones says it also raises an uncomfortable question for workers: what should happen to the time that has been freed?

Effort still carries moral weight

Jones points to psychology research on effort justification, which suggests people often place more value on outcomes that required more work. He also notes that many cultures treat hard work as a virtue, making tasks that feel easy seem less legitimate even when the result is useful.

That belief can collide with AI-assisted work. If a report, presentation or set of ideas takes a fraction of the usual time, the output may still be strong, but Jones argues that the worker may feel it was less earned.

The issue is also tied to professional identity. According to Jones, many people see a carefully written report or thoughtful analysis as evidence that they are capable and useful. When AI helps with structure, wording or analysis, the worker may start asking whether the finished product still counts as their own work.

Expertise shifts rather than disappears

The Conversation article argues that AI changes how competence is shown. In many jobs, expertise has been demonstrated through direct production: writing, analyzing and solving problems by hand. With AI, Jones says expertise may increasingly involve asking stronger questions, checking outputs, finding errors, adding context and taking responsibility for the final decision.

That shift does not make professional judgment less important, according to Jones. He says workers must still decide whether AI-assisted work is accurate, appropriate, ethical and useful.

Workplace norms may be slower to change. Jones argues that organizations may urge employees to use AI while continuing to reward visible effort and constant output. That leaves workers encouraged to be efficient while still feeling pressure to prove their value through busyness.

Saved time can become more work

The pressure may differ by role. Citing research on emotional labor, Jones writes that employees whose jobs depend on responsiveness, support and availability may find it especially hard to protect time saved by AI. For those workers, efficiency gains may turn into more unseen work rather than rest or reflection.

Jones warns that faster tools can reset expectations. A task that once took three hours may take 30 minutes, but the saved time can become capacity for additional assignments if managers do not define its purpose.

He argues that organizations should decide whether AI-created time will support better judgment, deeper thinking, collaboration, development or recovery. Otherwise, he says, productivity guilt will be reinforced by systems that use AI mainly to extract more output from the same workers.

The broader challenge, according to Jones, is to separate the legitimacy of work from the amount of visible effort behind it. In his view, the central test is whether the person using AI applied judgment, responsibility and care.

This story draws on original reporting from Phys.org.