AI Is Not Reducing Work. It Is Changing What We Do With Our Effort.

Adam Sturdee
Feb 17
4 min read

Recently on The AI Daily Brief, Nathaniel Whittemore referenced a study that caught my attention. Researchers followed roughly two hundred employees in a US technology company after they were given access to enterprise AI tools. The expectation, at least in theory, was that productivity gains might reduce workload. Instead, the opposite happened.

Link to the Harvard Business Review article

AI Doesn’t Reduce Work. It Intensifies It.

https://hbr.org/2026/02/ai-doesnt-reduce-work-it-intensifies-it

People worked more.

The research, later summarised in Harvard Business Review under the title AI Doesn’t Reduce Work. It Intensifies It., found that employees completed tasks faster but then filled the time they saved with additional work. They expanded the scope of what they were responsible for. They took on tasks that might previously have required another hire. Work began to spill into lunch breaks and evenings. It felt easier, so they did more.

That dynamic feels very familiar to me in schools.

When teachers first use AI tools to draft feedback, create resources or analyse a set of exam scripts, there is often a genuine sense of relief. The friction reduces. A worksheet that once took forty minutes now takes ten. A summary of a meeting can be produced instantly. A set of differentiated questions appears in seconds. The initial story we tell ourselves is that this will save time.

What I have observed, however, is that the time is rarely banked. It is reinvested.

Instead of producing one version of a worksheet, a teacher refines three. Instead of writing feedback for five books in a sitting, they push on to finish the whole class. A senior leader who can now draft a policy outline in minutes spends the saved time reworking language, checking alignment with other documents, and adding further detail. The standard rises because the cost of iteration falls.

This is not a criticism of teachers or leaders. It is a reflection of professional conscience.

In education, most people are wired to improve things when they see the opportunity. If a tool allows us to do something better, we tend to lean in. The Berkeley study suggests that the same psychology plays out in corporate settings. AI does not automatically create slack. It creates capacity, and capacity invites ambition.

There is also a subtler layer. When work feels cognitively lighter, it can be easier to justify extending it. Drafting alone can be draining. Drafting with AI support feels more collaborative, less effortful. That perceived reduction in effort can blur the boundary between necessary work and optional improvement. A ten minute task becomes a thirty minute optimisation exercise because it no longer feels punishing.

In schools, this has implications that go beyond individual workload.

If AI accelerates output but leaders quietly raise expectations in response, staff may experience intensification rather than relief. The technology becomes a ratchet. Reports become longer because they can be. Data analysis becomes deeper because it is now feasible. Communications become more polished. None of this is inherently wrong, but without conscious restraint, the system absorbs the efficiency gains and resets the baseline.

I have seen versions of this with transcript based lesson analysis. When teachers receive rapid feedback generated from a full lesson recording, they often engage more deeply with their practice. That is a positive shift. However, if the culture subtly moves from reflection to constant optimisation, the emotional load can increase. Faster insight can lead to more self scrutiny. More self scrutiny can lead to longer hours thinking about improvement.

The HBR article makes an important point. AI changes how work is experienced. It does not automatically change how much work exists. Organisations decide whether gains translate into reduced pressure or elevated standards.

In schools, that decision is leadership work.

If we genuinely want AI to reduce teacher workload, we have to protect the time it creates. That may mean explicitly stating that faster drafting does not mean more documents. It may mean capping iterations rather than perfecting them. It may mean resisting the temptation to turn every efficiency into an expanded initiative.

There is also a hopeful interpretation of the research. The fact that people voluntarily worked more suggests that AI can increase intrinsic motivation. When barriers drop, professionals feel more capable. They explore. They stretch. They improve processes that once felt too laborious. In education, that energy can be powerful if it is directed thoughtfully.

The danger is not AI itself. The danger is unexamined cultural drift.

If we assume that productivity gains will automatically translate into wellbeing gains, we will be disappointed. If we treat AI as a tool that must sit within a clear set of workload principles, we have a chance to shape its impact.

For me, the real question is not whether AI reduces work. It is whether we are disciplined enough as leaders to decide what happens to the time it frees.

In a profession already stretched, that is not a technical issue. It is a moral one.

Adam Sturdee is a senior leader and co-founder of Starlight, the UK’s teacher-first AI-powered transcript-based coaching platform for educators. His work sits at the intersection of dialogic practice, instructional leadership and responsible AI strategy for schools and trusts.

He will be presenting his research on AI-supported coaching at the BERA Teacher Education and Development Conference 2026: https://www.bera.ac.uk/conference/bera-tean-conference-2026

He is also speaking at the annual gathering of the SOPHIA Network – European Foundation for the Advancement of Philosophy with Children: https://www.sophianetwork.eu

AI Is Not Reducing Work. It Is Changing What We Do With Our Effort.

Recent Posts

Comments