AI was supposed to make work easier. Why are we busier than ever?

AI was supposed to lighten our workload. Why are we busier?

AI is often touted as the trump card to revitalize stagnant economies and boost the productivity of struggling companies. Indeed, an article published on Phys.org on March 23 highlights the current expectation that AI will address the UK's long-standing "productivity weakness." The UK government is also placing investment in AI and quantum technologies at the center of its growth strategy. However, the article's real question is not whether AI is useful, but rather a more fundamental question of why we are trying to make work more efficient in the first place.

The core of the article is clear. Productivity, or the measure of "how much output is produced per hour," is convenient for viewing the economy but does not directly reflect the value of work. It does not guarantee fair wages or stable employment. There is no assurance that work truly needed by society is correctly evaluated. In fact, focusing solely on efficiency can lead to cost-cutting and an over-reliance on tightly stretched supply chains, creating systems that appear smart but are actually fragile.

This issue is particularly visible in fields such as healthcare, caregiving, and education. The time nurses spend with patients, the time teachers adjust explanations based on students' understanding, and the attentiveness of caregivers to notice changes in the elderly—these jobs are defined by human interaction. Therefore, simply increasing speed, as on a production line, is not sufficient. As indicated by the economic concept of "Baumol's cost disease," such labor-intensive services are difficult to dramatically increase in productivity, yet they are indispensable to society.

Of course, the point here is not to say "AI is meaningless." In tasks like drafting documents, summarizing, assisting with coding, and setting up research entry points, AI is indeed strong. In February, Harvard Business Review argued that AI's promise is "burden reduction," but in reality, it tends to act in a way that increases work intensity rather than reducing it. Furthermore, an analysis by ActivTrak published in March showed that, based on 443 million hours of work data, instead of lightening the workload, the introduction of AI has accelerated the overall work process.

Symbolic of this is how time is used after AI adoption. The ActivTrak report introduced by Fortune showed that after AI was implemented, email time increased by 104%, chat and messaging by 145%, and the use of business management tools by 94%. Moreover, the average length of focused work sessions decreased by 9%, reducing the time available for deep thinking. Contrary to the ideal of AI taking over mundane tasks to create space, in reality, the "time that should have been freed up" is being absorbed by new requests, confirmations, corrections, and communications.

This paradox is also well reflected in reactions on social media. On Reddit, a comment suggesting that AI reduces the friction of individual tasks but increases the number of tasks and expectations gained support. Another user mentioned that AI has allowed them to venture into unfamiliar areas like front-end development, but this also means an "expansion of the defensive range." While being able to do more is progress, if it quickly turns into pressure to take on work that "didn't need to be done originally," the story changes.

A similar sentiment is visible on LinkedIn. Public posts repeatedly expressed the concern that "instead of reducing work because it got faster, it is seen as being able to do more because it got faster" and "the moment you show results, new expectations pile up." Some posts candidly stated that they don't want to show the efficiency gained from AI to others, as it would simply lead to more work being added. What is happening on the ground is not a backlash against the introduction of technology itself, but a distrust of the structure where the fruits of efficiency are collected as "additional work" rather than "rest."

Moreover, there is a difference in perception based on position. A Gallup survey from the fourth quarter of 2025 showed that the frequency of AI use in the workplace is higher among leaders, significantly ahead of individual contributors. Furthermore, a survey introduced by Fortune reported that executives are considerably more optimistic about the productivity-enhancing effects of AI than employees. From the top, AI is seen as a growth strategy, but from the ground, it can become a device that increases "work to process before thinking," and one must keep up with that speed.

On the other hand, not all expectations from management are mere fantasies. For example, Mark Cuban has expressed the view that AI agents could potentially shorten work hours by about an hour. However, there are conditions for that future to truly arrive. There needs to be institutional design to return the shortened time to rest and learning, a review of evaluation metrics, and a discussion on how the increased output from using AI will be distributed. Without these, AI will remain a tool for "working more" rather than a tool for "not having to work."

The importance of the Phys.org article lies not in dampening AI enthusiasm but in pointing out the danger of considering the future of work solely in terms of maximizing output. Being able to produce quickly is not the same as being able to work well. For work to support people, support society, and be compatible with life, "how much was produced in how many hours" is not enough. What AI is truly questioning is how much we are willing to sell our time and whose property the space created by efficiency will become.

Source URL Summary

Original article (published on Phys.org, reprinted from a contribution to The Conversation)
https://phys.org/news/2026-03-ai-boost-productivity-maximizing-output.html

Harvard Business Review article arguing that AI intensifies work rather than reducing it
https://hbr.org/2026/02/ai-doesnt-reduce-work-it-intensifies-it

ActivTrak summary page showing increased time in email, chat, and business management and decreased focus time after AI introduction
https://www.activtrak.com/resources/state-of-the-workplace/

ActivTrak news reporting key figures (443M hours analyzed, email +104%, chat +145%, etc.)
https://www.activtrak.com/news/state-of-the-workplace-ai-accelerating-work/

Inc. article introducing ActivTrak survey
https://www.inc.com/bruce-crumley/ai-is-boosting-productivity-but-data-shows-employee-workloads-are-getting-heavier/91316283

Fortune article introducing ActivTrak survey
https://fortune.com/2026/03/13/ai-isnt-reducing-workloads-its-straining-employees-time-spent-emailing-doubled-deep-focus-work-fell/

Gallup survey showing differences in AI usage frequency by position and deep usage in knowledge work and remote-capable jobs
https://www.gallup.com/workplace/701195/frequent-workplace-continued-rise.aspx

Fortune article reporting the gap in expectations for AI between executives and employees
https://fortune.com/2026/03/13/ceos-ai-mandate-employees-jobs-survey-nicholas-bloom/

Example of social media reaction 1: Reddit discussion on "AI reduces task friction but increases tasks and expectations"
https://www.reddit.com/r/datascience/comments/1r21ce9/new_study_finds_ai_may_be_leading_to_workload/

Example of social media reaction 2: LinkedIn post on "freed time becomes additional work, not free time"
https://www.linkedin.com/posts/georgettejupe_ai-doesnt-reduce-workit-intensifies-it-activity-7429516868815745024-PavA

Example of social media reaction 3: LinkedIn post on "showing efficiency leads to more work"
https://www.linkedin.com/posts/cshaeffer_ai-doesnt-reduce-workit-intensifies-it-activity-7440202483127406592-Oghl

Example of social media reaction 4: LinkedIn post on "AI doesn't reduce work on the plate, it raises expectations and cuts thinking time"
https://www.linkedin.com/posts/sean-m-white_ai-doesnt-reduce-workit-intensifies-it-activity-7427771653683867648-lnWB

Article reporting Mark Cuban's optimistic view that AI agents could shorten workdays as a contrasting example
https://www.businessinsider.com/mark-cuban-ai-agents-cut-workdays-hour-smart-companies-2026-3