Here's a longer, more flowing version written in a newspaper feature style, with fuller sentences and a more reflective tone throughout:

AI Was Supposed to Save Us Time. So Was Email.

There was a time, not that long ago, when office technology arrived carrying a simple promise: fewer hours lost to paperwork, fewer delays in communication, and fewer administrative burdens weighing down the working day.

Word processors would replace the typewriter. Email would replace the post. Information would move instantly instead of travelling in envelopes and filing cabinets. The friction would disappear and, in theory, so would the wasted time.

And yet, for many people, the opposite happened.

The typewriter disappeared, but the volume of writing exploded. The post tray emptied, but the inbox began to refill itself every few minutes. Instead of drafting one carefully considered letter, professionals now find themselves answering dozens, sometimes hundreds, of emails each day, often late into the evening. The very tools that were designed to remove friction from communication ended up multiplying it.

So when artificial intelligence arrived promising once again to take the pressure off knowledge workers by handling repetitive tasks, drafting documents, summarising meetings and generating ideas at speed, the pitch felt familiar. This time, we were told, machines would genuinely lighten the load. Humans would be freed to think, create and lead.

But early evidence suggests the reality may be more complicated.

The Research Behind the Optimism

A recent study published in Harvard Business Review followed approximately 200 employees at a US technology company over an eight-month period as they began using generative AI tools in their daily work. The tools were not imposed by management; they were made available and adopted voluntarily, allowing researchers to observe how they naturally changed workflows and habits.

What emerged was not a story of shrinking workloads or reclaimed hours. Instead, it was a story of acceleration.

Employees were able to complete tasks more quickly.

Drafts were produced faster. Analyses were turned around in less time. Routine elements of projects that once required sustained effort could be generated in seconds. But rather than translating into shorter working days or fewer responsibilities, these gains were absorbed almost immediately into increased output.

People did not work less. They worked more, and often at a quicker pace.

In some cases, they took on additional tasks simply because they now had the capability to do so. In others, the higher speed became an informal new benchmark, subtly resetting expectations about what could be delivered and how quickly.

The researchers described this dynamic as "work intensification," a phrase that neatly captures the tension at the heart of the AI productivity narrative. Efficiency at the level of a single task does not necessarily reduce the total amount of work done; it can expand it.

The Productivity Paradox, Again

The pattern is not new.

Each major leap in workplace technology has followed a similar arc. When spreadsheets arrived, financial modelling became faster and more detailed, which in turn led to more scenarios being analysed and more data being expected. When smartphones enabled constant connectivity, responsiveness became a cultural norm rather than an exception.

In each case, a tool that increased capability did not reduce effort overall. It raised the bar.

AI appears to be following that same trajectory.

If a marketing professional can now produce three campaign concepts in the time it once took to create one, the natural tendency is not to produce one and rest, but to explore three. If a manager can generate instant summaries of performance data, the temptation is to review more data more frequently. The tool expands what is possible, and what is possible gradually becomes what is expected.

Time saved quietly becomes time reallocated.

Why Work Expands Instead of Shrinks

There are structural reasons why AI's efficiency gains tend to translate into greater intensity rather than genuine relief.

First, AI lowers the cost of experimentation. Tasks that once felt too time-consuming to attempt now feel achievable. Employees broaden the scope of their roles, exploring adjacent responsibilities or taking on projects that would previously have been deferred. Individually, this feels empowering. Collectively, it increases workload density.

Second, AI does not remove the need for human judgement. Generated outputs must still be reviewed, edited, corrected and aligned with organisational standards. The nature of the work shifts from creating from scratch to supervising and refining, but the cognitive load remains. In fact, it can increase, as employees move rapidly between prompting, assessing and revising, fragmenting their attention.

Third, the availability of rapid output can erode natural stopping points. When it becomes easy to "just draft one more version" or "quickly check one more angle," the boundary between focused work and constant iteration begins to blur. The working day stretches subtly outward, not necessarily because anyone has demanded it, but because it feels efficient to continue.

The result is a paradox: productivity increases, yet so does pressure.

Efficiency on Paper, Strain in Practice

On paper, AI adoption often looks like a clear win. Output per hour rises. Turnaround times shorten. Teams appear more agile.

But the lived experience of workers tells a more nuanced story.

Employees in the HBR study reported working at a faster pace and, in some cases, beyond normal hours. The time saved by AI did not convert into leisure or deeper reflection; it converted into additional tasks. The day became denser, packed more tightly with deliverables and revisions.

This echoes what happened with email. Communication became instantaneous, which meant that responses were expected more quickly. The pace of exchange accelerated, and the volume grew. Rather than reducing correspondence, digital communication expanded it.

AI may be doing the same for cognitive labour.

A Cultural Question, Not a Technical One

None of this is an argument against artificial intelligence. The tools are undeniably powerful. They can remove repetitive manual processes, generate drafts that would otherwise consume hours, and unlock creative possibilities at speed. In fields ranging from marketing to software development to post-production, AI can automate elements that once required painstaking effort.

But technology alone does not determine whether work feels lighter or heavier.

If organisations treat AI as a way to increase output without adjusting expectations or redefining priorities, the technology will amplify existing pressures. If they instead use AI deliberately to eliminate low-value tasks and protect time for strategic thinking, the benefits may be more tangible.

The key distinction lies in how time savings are handled. Are they reinvested in more volume, or protected as space?

The Familiar Lesson

We have been here before.

IT did not eliminate paperwork; it digitised it and, in many cases, increased it. Email did not reduce correspondence; it multiplied it. Each wave of technology promised efficiency and delivered capability. What followed depended on how organisations chose to use that capability.

Artificial intelligence is unlikely to be different.

It can certainly make tasks faster. It can reduce friction in drafting, analysing and versioning. But unless businesses consciously convert those gains into genuine breathing room, AI risks becoming another accelerator layered onto already demanding workloads.

Technology creates capacity. Whether that capacity results in freedom or fatigue is a decision made not by the machine, but by the culture around it.