Why AI Isn't Simplifying Workloads as Promised

For many years, I have closely examined how digital technologies influence office-based professions. It's astonishing to reflect that my book Deep Work has now reached its tenth anniversary. Throughout this exploration, one recurring pattern has emerged consistently. A fresh technological innovatio

For many years, I have closely examined how digital technologies influence office-based professions. It's astonishing to reflect that my book Deep Work has now reached its tenth anniversary. Throughout this exploration, one recurring pattern has emerged consistently.

• A fresh technological innovation arrives, pledging to streamline those tedious elements of our professional responsibilities.
• Professionals eagerly anticipate the liberation of time, allowing greater focus on meaningful deep work and personal leisure activities.
• Instead, we discover ourselves even more overwhelmed with tasks, yet failing to generate additional high-impact results that truly advance key objectives.

This cycle repeated itself during the rollout of front-office IT systems, the widespread adoption of email, the rise of mobile computing devices, and more recently, the proliferation of video conferencing platforms.

Lately, I have grown concerned that artificial intelligence is following a strikingly similar trajectory.

These apprehensions were recently heightened by an insightful piece published in the Wall Street Journal, which bore the headline AI Isn’t Lightening Workloads. It’s Making Them More Intense.

The article draws on fresh data from ActivTrak, a software provider that scrutinized the digital behaviors of 164,000 employees spanning over 1,000 different organizations. The study's strength lies in its rigorous approach: it monitored specific AI users over a 180-day period both prior to and following their adoption of these tools, offering a precise view of the shifts that occurred. The findings paint a stark picture.

ActivTrak's analysis revealed that AI usage amplified activity in virtually every domain examined. For instance, the duration spent on email, instant messaging, and chat applications more than doubled. Similarly, engagement with business-management applications, including human resources platforms and accounting software, surged by 94 percent.

Remarkably, the sole area that did not experience this uptick was deep work: the intensive, uninterrupted concentration essential for tackling intricate challenges, developing formulas, crafting content, and engaging in strategic planning. Among AI users, time allocated to such focused efforts declined by 9 percent, while non-users showed virtually no variation.

This outcome represents a troubling scenario where individuals labor at a quicker pace and with greater exertion, but primarily on superficial, cognitively draining activities. These tasks demand constant context-switching, providing only marginal contributions to organizational goals when contrasted with the more demanding pursuits of deep work.

The precise reasons behind AI's role in this phenomenon remain somewhat elusive. However, a compelling hint emerges from Berkeley professor Aruna Ranganathan, featured in the article. She observes that AI renders extra tasks seem straightforward and readily approachable, thereby fostering a sensation of forward momentum.

This observation echoes the historical parallel with the debut of email communication. Undoubtedly, composing and dispatching emails proved far more efficient than navigating fax machines or cumbersome voicemail systems. Yet, once low-effort messaging became available, employees restructured their entire workflows around relentless exchanges of rapid-fire communications. This frenzy conveyed an illusion of productivity in a vague, motion-oriented manner, but it ultimately undermined nearly every other dimension of their roles and left workers broadly dissatisfied.

Contemporary AI applications appear to be perpetuating a comparable pattern, but applied to bite-sized, standalone assignments. Professionals now engage in rapid, iterative dialogues with chatbots, continuously polishing textual outputs and producing preliminary versions of reports and presentations that frequently fall short of practical utility due to their rough quality. For those with advanced technical skills, this might extend to overseeing clusters of AI agents working in tandem to accelerate these processes even more. From this vantage, everything appears industrious because individual operations seem to accelerate, and overall busyness escalates noticeably.

Yet, we must question whether these accelerations are targeting the most critical components of our professional duties. Are we truly enhancing the elements that deliver the greatest value, or are we merely amplifying superficial busyness at the expense of substantive progress?

In reflecting on these dynamics, it's evident that technology's promise of efficiency often morphs into an unintended trap of heightened frenzy. The ActivTrak data underscores how AI, despite its capabilities, is channeling efforts toward low-level interactions rather than elevating the quality of output. This misalignment could stem from the ease with which AI handles routine tasks, tempting users to pile on more of them instead of protecting time for reflection and creation. Historically, each wave of office tech has followed this path, from email's endless threads to the perpetual pings of Slack and now the hypnotic pull of generative AI interfaces. To counteract this, workers might need deliberate strategies, such as time-blocking for deep work or imposing strict quotas on AI-assisted communications, ensuring that technological aids serve higher purposes rather than dictating the rhythm of the workday.