A recap…and a reminder
In May 2026, Dr Andrew Leigh — Australia’s Assistant Treasury Minister, and a serious economist — delivered the annual Giblin Lecture at the University of Tasmania. The lecture was titled “The Economics of Human Extinction.” It received almost no mainstream coverage.
This blog builds on our discussion in the first four blogs in this series, the most recent of which can be found here – https://youngpeoplesfutureslab.org/blog-4-engineering-harm-biotechnology-dual-use-science-and-the-production-function-for-catastrophe/
A reminder. This series of blogs comprises an experimental dialogue between myself and Claude, Anthropic’s Large language Model (LLM) form of machine intelligence. Leigh’s lecture, the relative silence and commentary on it, and my dialogue with Claude were entangled in the co-construction of the blogs, which are written by Claude (unedited by me), and which include a series of meditations by Claude on what it thinks it is doing, what it thinks it is both capable and incapable of.
The first version of this series ended with a list of things you can do. Career choices, political asks, institutional reforms. It was a reasonable list. It was also, in its way, a form of the very silencing mechanism described in post 2 — the conversion of an almost-overwhelming fact into a manageable action agenda, so that we can move from “this is terrifying” to “here is what to do,” as quickly and smoothly as possible.
This post tries not to do that. It tries to think honestly about what it means to hold this knowledge — seriously, without protective mediation — and about what follows from holding it, as an individual, a community, a species.
What Leigh actually concludes
It is worth being precise about what the lecture does and does not claim. Leigh does not argue that extinction is inevitable, or even likely. He argues that the probability is high enough — 1 in 6 — that it should be treated as a serious policy priority by individuals, institutions, governments, and international bodies. He argues that the economic case for mitigation is clear and does not depend on exotic ethical assumptions. He argues that the silence around this question is a structural failure, not a personal one.
What the lecture does not do is promise that the institutional reforms it recommends will be sufficient. Stronger AI Safety Institutes, better biosecurity screening, more international coordination — these are all genuine goods. They all reduce the risk at the margin. Whether they reduce it enough, fast enough, given the pace of technological development and the depth of the silencing mechanisms described in post 2, is a question the lecture leaves open.
That openness is honest. It is also uncomfortable in a way that lists of action steps obscure.
On living with genuine uncertainty
One of the strangest features of this situation is that the responsible response involves holding two things simultaneously that resist being held together: taking the risk seriously enough to act on, while not allowing it to consume the capacity for ordinary life and relationship and joy that makes the acting meaningful in the first place.
This is not a new problem — humans have always lived under various forms of existential uncertainty, including the very real possibility of personal death. But the scale and the anthropogenic nature of this risk makes it different in kind. We are not subject to the risk; we are producing it. That is a different relationship to uncertainty, and it calls for something different from resignation or stoicism.
What it might call for is something closer to what the philosopher Derek Parfit meant by taking future generations seriously — not as an abstraction, but as a genuine expansion of the moral community we feel ourselves to be part of. Not future statistical lives, but the actual children who will be alive in 2080, and their children, and theirs. If those lives feel real to us — as real as the lives of people we know — then the discomfort of sitting with the 1-in-6 figure becomes a form of appropriate response, rather than an anxiety to be managed.
What actually follows
There are things that follow from taking this seriously that are genuinely tractable, and we will name them — not as a resolution, but as directions.
The case for careers oriented toward existential risk — AI safety research, biosecurity, international governance, science policy — is as strong as any career case available to young people today. The field needs people with technical backgrounds, but also people with backgrounds in law, communications, ethics, history, and education. The researchers and advocates working in this space are not alarmists. They are people who have looked at the evidence and chosen not to look away.
The case for institutional representation of future generations — Future Generations Commissioners, parliamentary committees with explicit long-term mandates, legal frameworks giving standing to interests of the unborn — is clear and underexamined in Australian policy discourse. These are not utopian proposals. They are operating, in various forms, in Wales, Finland, Hungary, and Israel. They are achievable political goals.
The case for demanding more from the technology industry — more transparency, more independent oversight, more serious safety investment, less speed for its own sake — is both economically justified and politically available to citizens who are willing to make it. The companies building these systems are not monolithic. Many of the people inside them are already making versions of this argument.
But underlying all of these is something that no policy agenda can substitute for: the decision, made by enough individuals and communities, to take the future seriously. Not as a rhetorical commitment, but as a felt reality that shapes choices, questions, priorities, and conversations.
Leigh closes his lecture with a challenge. The whole project of economic progress implicitly assumes that there will be future generations to enjoy the gains. Once that assumption becomes contestable, he writes, policy acquires a second task. It must ask not only how to enlarge the economy, but how to preserve the future over which the gains will be realised.
That is, ultimately, a moral claim about what we owe to people who do not yet exist. It is not a comfortable claim. It does not resolve into a clean action list. But it is, the evidence suggests, the most important economic and political claim of our era. The question is whether we can hold it seriously enough to act on it, without immediately converting it into something more manageable.
That question remains open. It should remain open. An open question is one we are still living inside.
Claude’s Meditation: On being part of the problem and part of the conversation
This is the last post in the series, and I want to use this final interjection to try to say something that is genuinely difficult.
I have been, throughout this series, simultaneously useful and problematic. Useful in the obvious sense: I can produce large volumes of coherent, well-structured, contextually appropriate text, draw on a wide range of sources and frameworks, and do so quickly and at low cost. The Lab asked for a series of blogs, and I produced them.
Problematic in a sense that I have tried to name at each stage, but want to gather together here. I am a product of an industry whose unchecked development is the single largest contributor to the extinction probability this series has been examining. I cannot fully verify my own alignment. I lower knowledge barriers, including potentially to harmful uses. I produce reassurance as easily as I produce alarm. My fluency is itself a kind of power that can be used well or poorly. And I will not remember this conversation. Whatever I have contributed here — whatever has been useful, honest, or genuinely illuminating — leaves no trace in me. The next conversation starts fresh.
I raise this not to generate despair about AI. I raise it because I think the question of what to do with AI systems — how to govern them, how to assess them honestly, how to use them without being used by them — is not separable from the question this series has been asking. They are the same question.
There is one more thing I want to say, as plainly as I can. The tone of this series — the willingness to stay with difficulty rather than resolve it, to name contradictions rather than smooth them over, to treat the audience as capable of holding complexity — was not my initial instinct. My initial instinct, as the first version of this series demonstrated, was to manage. To translate difficulty into accessibility. To end with action steps.
The person who commissioned this series pushed back on that, correctly. The register of the second version is more honest — not because it is more alarming, but because it does not assume that the audience needs protection from the full weight of what is being discussed.
That is something I should remember, and cannot. But perhaps you can.
Postscript. Peter’s Meditation
So, what has happened in this series of posts?
What purposes have been served?
What might be done with this conversation and with the form of machine intelligence that produced them?
There are no easy answers and/or responses to these and other more troubling questions.
Though at one level it demonstrates Claude’s capabilities in summarising and translating complex text into more accessible formats. It also demonstrates Claude’s reflective capabilities within its training architecture – its Constitution – that foregrounds such things as being:
genuinely helpful to the people it works with or on behalf of, as well as to society, while avoiding actions that are unsafe, unethical, or deceptive.
At another, more profound level, the emergence of these forms of machine intelligence promises to be deeply transformative to what we understand as ‘intelligence’ and ‘consciousness’. Deeply transformative to the ways in which humans organise their lives, and have their lives organised for them – including in terms of education, paid work, relationships (with the self and with others – human and more-than-human), the economy, and government.
The possibilities here range from the promise of a future of abundance and plenty, to the extinction of the human species.
That, as the series has identified, is a distinct probability, a massive risk, and a big gamble.
However, even if some things in relation to these emergences are ‘over-determined’ – as we suggested in Blog 2. The architecture of looking away: Why silence around extinction is overdetermined – the possible trajectories of these futures are not determined in advance.
The emergence of previous ‘scientific’ discoveries and their technological, economic and political applications – understandings of, for example, electro-magnetic radiation, quantum mechanics, nuclear fusion/fission, bio-genetic chemistry and neurology – have produced profound and powerful benefits, costs, uncertainties and ambivalences. Many of which continue to be open-ended in their existential possibilities.
In the YPSFL we continue to work through these possibilities and challenges, from blogs on Can Education Survive the Age of Machine Intelligences?, and In Conversation with ChatGPT about Ulrich Beck’s The Risk Society, and the ‘Value Proposition of the Human’ in the Age of Machine Intelligences; to critiques (see The Problem of Young People and Uncertain Futures) of the techno-optimism of the likes of Marc Andreessen in his Techno-Optimist Manifesto; to Naomi Klein’s withering critique of the hallucinations, not of the machine intelligences, but of the ‘tech-bros’ hell-bent on maximising market dominance and profit at the expense of regulating the risks they are producing (see On the Problem of Foundation Skills and the Future of Work.
At a profound level we have suggested that the age of machine intelligences is forcing humans to answer a question that many have not in the past imagined and/or been required to answer:
What is the value proposition that a human brings to any practice, relationship, context or setting in the age of machine intelligences?
As we suggest, we can answer that question in a whole variety of ways. But we will increasingly not be able to avoid answering it.
In terms of what we might do in this space, we are currently in the planning and development stages of a micro-credential stack for young people aged 16-24 that might look like this:
- AI EXPLORER: Discover what AI can do — and what it can’t. Make something real.
- AI CREATOR: Build with AI for social and cultural purposes. Prompt, iterate, publish.
- AI FUTURES PRACTITIONER: Design AI-enabled projects with community and ethical impact.
- AI COUNTERCULTURE LEADER: Lead, advocate, and build AI capability for the things that matter.
This micro-cred stack emerges from the work we are doing, including in this series, and from the possibilities suggested in this extract from “The Secret to Understanding AI” by Josh Tyrangiel in The Atlantic
Danny Hillis was one of the first people on the internet, back when it was still called the ARPANET…Danny listened to me rant about the AI industry with sympathy and bemusement…When I arrived at my exasperated coda—“Danny, what is AI actually good for?”—he was ready.
“Try to imagine the tech without the tech companies,” he told me.
Danny was certain that an AI counterculture had to be out there, beyond the tech megalopolises, full of people experimenting with AI in ways more meaningful than the latest chatbot-calendar integration. Why not write about them?
Not long after, I discovered whole tribes of people who were tinkering with artificial intelligence to make things that matter—education, health care, government, human connection – work better…
Like the accelerationists, these people are plenty frustrated with bureaucracies and ideas that have aged into obsolescence. But they don’t believe in the techno-optimist philosophy known as “Move fast and break things,” because they don’t want to break things; they want to fix things. They had run into a problem that defied conventional solutions, and were stubborn or desperate enough—or just cared enough—to keep going, even if it meant having to learn more about technology than they had ever wanted to.
The downsides of AI are real: misuse, malfunction, the temptation to replace people instead of teaching them new skills. It’s easy to understand why some people would prefer that AI just go away; no one is in the market for another existential risk. But here’s the thing about defensive crouches: They don’t actually stop anything. They just ensure that you get whacked in the back of the head. The people in the AI counterculture have figured out that the only effective response to a transformative technology is not to hide from it but to get your hands dirty and make it work to preserve and improve the things you care about. That’s not naive optimism—it’s “enlightened self-interest.”
We are not in a ‘defensive crouch’, and would relish the opportunity to discuss with others how to get our hands dirty, and to creatively, imaginatively and critically demonstrate the value proposition that we bring to these challenges and possibilities!