The documentary The AI Doc: Or How I Became an Apocaloptimist asks one of the defining questions of our time: Will artificial intelligence save humanity or destroy it? But perhaps we should be asking a different question. Who will control this new intelligence, and why are we allowing the people developing the technology to define our understanding of the future?
There is something peculiar about the way we discuss artificial intelligence today. We are developing technology that is already changing how people work, communicate, create, and solve problems. Yet much of the public debate revolves around two almost religious visions of the future. Either AI will liberate humanity from disease, poverty, and work, or it will evolve into a superior intelligence that ultimately renders human beings obsolete. Somewhere between these extremes lies the reality in which we actually live, but it receives surprisingly little attention.
This is what makes the 2026 documentary The AI Doc: Or How I Became an Apocaloptimist, directed by Daniel Roher and Charlie Tyrell, so interesting. It attempts to make sense of a technological shift with few obvious historical parallels, at least in terms of its speed and potential reach. But the film also illustrates a larger problem with the entire AI debate. We have become so fascinated by what machines might one day become that we risk overlooking what the people behind them are already changing.
an yet verify.
When Humanity's Future Is Seen Through the Birth of a Child
The film begins with something deeply human. Daniel Roher is about to become a father, and he wants to understand the world his child will grow up in. His anxiety is understandable. We are living through a technological transformation so rapid that even those working at its centre struggle to offer credible predictions about what AI will look like in five or ten years. Roher therefore seeks out some of the industry's most influential figures, including Sam Altman of OpenAI, Dario Amodei of Anthropic, and Demis Hassabis of Google DeepMind. He also meets researchers, critics, and thinkers who warn that the consequences could extend far beyond anything we can currently comprehend.
The personal framing is effective because it shifts the discussion away from technical specifications and towards a fundamental human concern. What does the future mean when you are preparing to hand the world over to the next generation? Few parents would dismiss that question. Yet there is a weakness here that extends well beyond the film itself. Once AI is presented primarily as a question of humanity's survival, almost everything else becomes secondary. Employment, education, economics, democracy, and the distribution of power can start to look like minor details when the alternative is the possible extinction of our species.
But we must be able to hold two ideas in our heads at once. It is reasonable to investigate the possibility that highly advanced AI systems might one day become difficult to control. It is not reasonable to treat that outcome as a scientifically established certainty. The International AI Safety Report published in February 2026, drawing on the work of more than a hundred experts, recognises both the rapid development of AI capabilities and the risks that can already be documented. It also emphasises the considerable uncertainty surrounding forecasts of future capabilities and the possibility of losing control over these systems.
That distinction matters. The fact that a risk is possible does not mean we know how likely it is. And uncertainty does not mean we are entitled to ignore it. A serious debate must accommodate both insights without collapsing into either apocalyptic propaganda or uncritical technological optimism.
Perhaps that is why the film's central concept, apocaloptimism, is so attractive. It offers an intellectual middle ground: acknowledge even the darkest scenarios while continuing to believe in a better future. But positioning oneself halfway between two extremes is not automatically a sign of wisdom. Sometimes one side has stronger evidence. Sometimes both rely on assumptions that nobody can yet verify.
The People Building the Future Also Get to Tell Us What It Means
Perhaps the most interesting aspect of the documentary is not its visions of the future, but the people invited to describe them. This creates a journalistic problem that deserves far more attention. The leaders of the major AI companies are, of course, important interview subjects. They possess knowledge, resources, and experience unavailable to most people. But they are also executives operating within one of the most capital-intensive and fiercely competitive industries of our time.
Sam Altman, Dario Amodei, and Demis Hassabis are not detached observers looking at an uncertain future from neutral vantage points. They lead organisations that are actively creating that future. Their statements must therefore be understood both as technical assessments and as expressions made within a particular economic and strategic reality. That does not make their arguments wrong. It makes critical follow-up questions essential.
This is also where several reviewers have criticised the documentary. Among others, The Verge and Screen Daily have questioned how far the film actually goes in scrutinising the powerful figures it interviews. Access to those in power is impressive, but access and scrutiny are not the same thing. A journalist can secure an exclusive interview with one of the world's most influential business leaders and still fail to ask the most important questions.
I would like to know, for example, why these companies believe they should wield such decisive influence over humanity's technological future. What democratic mandate do they have? Who oversees how their models are trained and deployed? What happens when commercial incentives clash with safety? And who will be held accountable if the technology causes widespread harm?
These questions tend to disappear when the conversation turns instead to whether AI will create paradise or destroy humanity.
We have seen this pattern before. Social media was supposed to bring people closer together, democratise information, and give everyone a voice. Much of that promise was fulfilled. Yet the same platforms also developed some of the world's most sophisticated systems for behavioural analysis, attention manipulation, and the commercial exploitation of personal data. The technology's opportunities were real, but so were its business models. And the interests of users were not always what determined how those systems evolved.
It would be naive to assume that the AI industry will automatically behave differently.
The Catastrophe Does Not Require Superintelligence
One of my strongest objections to the prevailing AI debate is that it so often turns the most dramatic future risks into the main attraction, while profound social changes are already happening in front of us. We do not need to create a conscious superintelligence to transform the labour market, concentrate economic power, or make it harder for people to distinguish authentic information from manipulation.
Today's AI systems can already automate or improve the efficiency of parts of highly skilled work. According to a 2025 analysis by the International Labour Organization, roughly one in four employed people worldwide work in occupations that are exposed to generative AI to some degree. At the same time, the organisation stresses that the transformation of tasks is a more likely overall outcome than the disappearance of entire occupations. That is an important qualification in a debate where both mass unemployment and complete liberation from work are sometimes presented as virtually inevitable.
But even if jobs change rather than disappear, an important question remains—one that technology cannot answer for us. Who benefits from the increased productivity?
If one person using AI can perform work that previously required several people, the economic opportunity is enormous. A company can produce more, reduce costs, lower prices, improve working conditions, or increase profits. There is no law of nature ensuring that the gains will be shared fairly. The Industrial Revolution taught us that technological productivity and social progress are not the same thing. Shorter working hours, safer workplaces, and greater economic security did not emerge simply because machines became more efficient. They required political decisions, social organisation, and struggles over how the new wealth would be distributed.
Why should the AI revolution be any different?
Meanwhile, a vast physical infrastructure is expanding, though it rarely receives the same attention as grand visions of future intelligence. AI is not some magical digital force that exists independently of the material world. It depends on advanced semiconductors, enormous data centres, electricity, cooling, capital, and functioning supply chains. According to the International Energy Agency's updated 2026 assessment, global data-centre electricity consumption could rise from approximately 485 terawatt-hours in 2025 to around 950 terawatt-hours in 2030.
That development raises very concrete questions about energy supply, investment, and the resources of local communities. At the same time, it is important to recognise that AI may also contribute to more efficient energy systems and technological innovation. The point is to acknowledge both the costs and the potential benefits rather than portraying the technology as either an environmental disaster or an automatic solution.
The real question about the future is therefore not simply how intelligent these systems may become. It is also how much power we are prepared to concentrate in the companies and governments that control their infrastructure.
The New Paradox: More Power for the Giants, More Freedom for Individuals
There is, however, another side to AI development that I believe receives far too little attention. While the world's largest technology companies are investing staggering sums in data centres, models, and infrastructure, individuals are gaining access to tools that were once available only to large organisations.
This may be the most fascinating contradiction of the entire transformation.
A person with an idea, some experience, and access to modern AI tools can now build working digital prototypes, analyse complex problems, process information, and experiment with technical solutions without first having to establish an organisation. I see this as something far more significant than another improvement in productivity. The distance between having an idea and turning it into something tangible is beginning to change.
For people who have spent years working with ideas and projects, the difference is striking. In the past, moving from an idea to a functioning result could require substantial funding, specialists, coordination, and lengthy development processes. Today, an initial prototype can sometimes be built in a matter of hours. It is not necessarily finished, secure, or even particularly good. But it exists. It can be demonstrated, tested, challenged, and improved. That difference is fundamental.
This does not mean that everyone suddenly becomes a skilled programmer, researcher, or entrepreneur. Experience, judgement, subject knowledge, and the ability to understand people's real needs become more important, not less. AI can generate an impressive amount of code without understanding whether the resulting product is actually needed. It can produce confident-sounding analysis without the reasoning holding up. This is where human responsibility must begin, not end.
But even this new freedom comes with dependencies. Independent creators using AI often rely on models, cloud services, and platforms controlled by the very same large companies whose power we are discussing. A change in subscription terms, new restrictions, or the closure of a service can affect their ability to continue working. The democratisation of creation does not therefore automatically mean the democratisation of technological infrastructure.
AI may give rise to a new generation of independent inventors and small digital businesses. It is also possible that a growing share of the value they create will be captured by the platforms on which they depend. Which tendency will dominate remains an open question.
We Should Fear the Future Less and Scrutinise the Present More
What makes The AI Doc relevant is that it captures a genuine uncertainty. Nobody knows exactly where AI development is heading, not even the people closest to it. The film's mixture of anxiety and optimism therefore reflects something real. Yet a documentary that seeks to understand perhaps the most important technological transformation of our time must dare to do more than place competing visions of the future side by side.
We need to be able to discuss long-term existential risks without presenting them as inevitable. We need to acknowledge AI's extraordinary potential without treating corporate promises as scientific findings. And we need to question the concentration of power and the underlying business models without denying that the technology is already giving people new opportunities to create, learn, and solve problems.
Above all, we need to stop talking about the AI revolution as though it were a weather system approaching from the horizon. It is not something that simply happens to us. It is shaped by people, companies, institutions, economic interests, and political decisions. That also means we still have opportunities to influence its direction.
I believe AI can become one of humanity's most significant tools. I can already see how the technology is changing the conditions for independent creation and making things possible that were previously beyond practical or financial reach. But that is precisely why I become sceptical when the discussion is reduced to a choice between salvation and destruction. It risks turning us into spectators of a future that we still have the power to shape.
Perhaps the most disturbing possibility is not that machines will one day become more intelligent than human beings. It is that we are already handing decisive parts of the future's power and resources to a small number of actors while debating what might happen in ten or twenty years.
By the time we recognise the difference, the most important decisions may already have been made.
By Chris...
Instadocs: AI Gone Wild | Official Teaser | Netflix