AI Utopia, Well Maybe?

We all wonder about the future. Will our lives become easier? Will we have more free time? Will we be happier? These questions aren't new, but right now they feel more urgent, because of the emerging technological force that could actually change the answers.

The Dream Is Not Unreasonable

Imagine a working week, not too far from now, where you check in on your AI agents over a coffee on Monday morning, point them at the week's priorities, before spending some time on things that actually require being a human. A conversation with a client. A strategic decision. A creative problem that needs judgement and taste. Then after working three hours each day you finish at lunchtime on Thursday for a nice long weekend.

This is not complete science fiction. A shorter work-week could potentially be introduced by the AI labs, in a similar way that Henry Ford introduced the five day work week. The difference with AI is the pace of change. Previous waves of automation were gradual enough that society had time, however imperfectly, to absorb the disruption. What we’re dealing with now feels like something quite different, and that changes everything about how the transition might play out.

The Messy Present

John Maynard Keynes predicted in 1930 that his grandchildren would work a fifteen-hour week, freed by the compounding gains of technology ("Economic Possibilities for our Grandchildren", 1930). He was right about the productivity gains. He was plain wrong about where the time savings would go. And so far with AI, we are too.

The honest reality today is that AI is not letting us work less. For knowledge workers who have adopted these new tools enthusiastically, it is making us work more. Research by Aruna Ranganathan and Xingqi Maggie Ye of UC Berkeley's Haas School of Business, published in Harvard Business Review in February 2026, tracked 200 employees at a technology company for eight months and found that AI tool adoption consistently intensified work rather than reducing it. The hours AI saves get absorbed immediately by higher output expectations. The productivity gains are real; but they are not translating into free time.

There is also a more subtle problem. When AI handles the routine parts of work, what remains is relentless concentration and high-stakes decision-making, the unglamorous grind and cognitive overload of keeping agents running, reviewing what they produce, correcting errors and maintaining pipelines. The small tasks that once gave the brain a quiet moment to recover are disappearing. The result is a working day that feels more exhausting than before, even when the output is greater. And anyone who has spent an afternoon wading through AI-generated content knows it is not the leisurely future we promised ourselves.

But it is worth asking whether we’re drawing the wrong conclusion from this exhaustion. The railways did not deliver their benefits the moment the first locomotive was fired up. They required an enormous prior investment of time, capital and human labour, to lay thousands of miles of track, build stations, and figure out how to run a schedule. The productivity and freedom came later, once the infrastructure was in place. We may be in the equivalent moment with AI: not failing to realise its promise, but laying the tracks. Our workflows are not yet optimised, our agents are not yet fully-configured, and the guardrails are still being built while we are also trying to do our existing jobs. The exhaustion of this period may be the exhaustion of construction, rather than the exhaustion of a broken system. Once the infrastructure matures, and the right automations are in place, the workload may finally start to reduce.

The Distribution Problem

Even if AI eventually creates vast wealth and frees up enormous amounts of human time, there are many unresolved questions: who benefits, and when? The Industrial Revolution created vast wealth, but it took a century of political struggle before that wealth was broadly distributed. The digital revolution has been similarly uneven. Discussions about Universal Basic Income (UBI) and profit‑sharing mandates are well-intentioned, but the political, economic and societal challenges are daunting, and the debate has barely moved beyond the theoretical. The transition will be uneven across geographies, industries and income levels, and the support structures people need will likely not arrive in sync with the changes imposed on the labour market.

A Generation Waiting on the Sidelines

There is one group for whom the messy transition carries a particular risk, and it may not be the one most people would predict. The assumption is that young people will adapt most naturally to AI. They are digital natives, comfortable with new tools. Surely they are the winners here?

Comfort with technology is not the same as having the domain knowledge and experience to supervise it effectively. Entry level positions have always served a deeper purpose than simply performing basic, repeatable tasks. They provide the environment to build domain knowledge and judgement, how to make ethical decisions, and how to communicate and collaborate with others.

As AI automates the entry-level tasks that provided that foundation, the bottom rung of the career ladder is disappearing. Young people are left in professional limbo, finding it harder to get a foot in the door, since AI can do the basic tasks, and they are not yet qualified or experienced enough to provide sufficient oversight of AI. This could be a lost generation in the making, not through dramatic displacement, but through a quiet stalling.

The response cannot simply be to tell young people to learn AI tools and concepts. What they actually need are the skills that make working with AI genuinely valuable: critical thinking, ethical reasoning, and clear communication. The problem is that most education systems are not yet teaching these topics systematically, and the workplace, which traditionally filled that gap, is now precisely where the gap is widest. I have written separately about the case for shifting these skills into education before young people reach the job market: Workplace Readiness Needs to Shift-Left.

The Governance Gap

Layered on top of this is a challenge that rarely features in optimistic projections: the world has not agreed on how to manage any of this. AI governance is a patchwork of the EU's AI Act, American executive orders, and China's own frameworks shaped by its own priorities. International summits in Bletchley, Seoul, Paris and New Delhi have generated shared language but no binding agreements and no enforcement mechanisms.

Chatham House concluded in early 2026 that meaningful global AI governance may only become politically feasible in the event of a crisis. The United States and China are racing to achieve dominance in AI, and neither has a strong incentive to constrain itself in the name of global coordination. The infrastructure for managing a technology of this significance simply does not exist.

Why Optimism Still Makes Sense

Taken together, that is a lot to hold at once: burnout in the present, unresolved distribution questions, a generation at risk, and a global governance vacuum. It is easy to lean towards a pessimistic viewpoint. But technology has always confounded pessimists, and sometimes in ways that seem almost comical in retrospect. When trains were invented in the early nineteenth century, some physicians warned that the unprecedented speed and acceleration would crush passengers' internal organs. The fear was real, and the technology was genuinely new and strange, but the critics were wrong about what it would do for humanity. It connected people, opened up economies, and shrank the world in ways that were overwhelmingly positive. You might say that this technology transformation is different, but they were all different in the past too.

As AI emerges from the trough of disillusionment it will hopefully follow a similar path. The concerns are not irrational, but the typical path of technological evolution, however messy, has mostly evolved towards greater prosperity and human capability. Unlike previous technologies, AI holds the potential to help solve bigger problems that have stumped us for decades: accelerating drug discovery, improving education, addressing climate change and making expertise more widely accessible. The potential upsides are significant.

The governance gap is real, but it is being noticed and governments are starting to wake up to the risks as more powerful AI models are being released. The history of international cooperation is not a story of elegant solutions arrived at smoothly, it’s a story of problems eventually being addressed when the pressure exceeds an undefined threshold. And on the question of work and free time, we simply haven’t had time to adapt yet. The conversations about sharing AI generated productivity and financial gains with workers are happening now in ways they were not a few years ago.

The Gap Between Here and There

The hardest part of being an optimist about AI is not believing the destination is good, it is holding that belief while being honest about the distance between here and there. The transition will be messy. The AI governance and alignment questions are not even close to being resolved. The wealth distribution problem will not resolve itself without deliberate political leadership that is unlikely to happen quickly, if at all. And a generation of young people risk being stalled at the starting line.

But the dream of a future with more free time, more connection, more human freedom is not naive. It is actually quite plausible, if we are willing to do the hard work of building the social, political and institutional structures that will allow it to arrive equitably.

For now, a leisurely coffee with friends in the afternoon is still aspirational. Most of us are reviewing agent outputs and wondering where the afternoon went. That is the honest place we are in, and it’s worth acknowledging, even for those of us who remain convinced it is not where we will end up. As for a four day work week, we can only live in hope!

Further Reading

• Workplace Readiness Needs to Shift-Left
• The AI Adoption Gap: Why Regulated Companies Can't Move As Fast As The Hype