This thesis is the basis of the public lecture Proof of Human: AI, Copyright, and the Fight for Creative Authorship, delivered at SISTER in Manchester's Innovation District (18–19 May 2026) as part of Creative Manchester's programme at the University of Manchester, followed by a conversation with John McGrath, Artistic Director & Chief Executive of Factory International. It is also the basis of the forthcoming book of the same name from Manchester University Press.
I. The Second Enclosure
Between the fifteenth and nineteenth centuries, the commons of England were systematically fenced. Land that had sustained communities for generations was privatised, enclosed, and extracted from. The peasantry were not consulted. They were displaced, proletarianised, and told this was progress. The Enclosure Acts created the conditions for industrial capitalism and for the social upheavals that followed.
We are living through a second enclosure. But this time, the commons being fenced is not land. It is the human mind itself: our attention, our creativity, our behavioural patterns, our relationships, our labour, our identity. The new lords are not aristocrats with parliamentary influence. They are platform owners, data brokers, and the architects of artificial intelligence systems trained on the sum of human expression — without consent, without compensation, without acknowledgment.
This thesis does not argue that a coordinated conspiracy is underway, nor that technological development is intrinsically malign. It argues that incentive-aligned systems can produce civilisational outcomes independent of intent, and that the absence of countervailing infrastructure allows extraction to harden into enclosure. The question is not whether powerful actors are good or evil. The question is whether the structures they operate within produce outcomes compatible with human flourishing — and what infrastructure might be built to ensure they do.
This thesis argues that we are witnessing not merely technological disruption, but a fundamental restructuring of human agency and economic participation. The mechanisms are technical. The consequences are civilisational. And the window to build countervailing infrastructure is narrowing.
II. The Accelerationist Mechanism
There is a pattern to the crises of our era. Each shock — financial, epidemiological, geopolitical — accelerates the same set of outcomes: wealth concentration, institutional erosion, surveillance normalisation, democratic retreat. The beneficiaries are consistent. The losers are predictable. What varies is only the proximate cause.
Whether these outcomes are the product of deliberate coordination or emergent behaviour among actors with aligned incentives is, in most respects, immaterial. The effects are structurally identical. Capital flows to jurisdictions beyond democratic accountability. Public services are starved, privatised, or contracted to surveillance-adjacent firms. Labour is casualised, automated, and stripped of bargaining power. The social contract frays — and this fraying encounters no countervailing force from those with power, because the fraying aligns with their material interests.
The COVID-19 pandemic accelerated these dynamics with remarkable efficiency. Whatever its origins, the crisis functioned as a stress test that forced latent beliefs into the open. Platform-level content moderation systems flagged, demoted, and removed accounts at unprecedented scale. Metadata capture expanded. Behavioural classification systems — developed for advertising — were repurposed for identifying patterns of non-compliance and network mapping of ideological affiliation.
Concurrently, law enforcement agencies in Europe and the UK successfully compromised encrypted communications platforms used by organised crime networks. Operation Venetic, based on the infiltration of EncroChat, led to over 2,600 arrests in the UK alone. The Sky ECC takedown yielded similar results across the continent. These operations were legitimate law enforcement successes — but they also normalised the principle that secure communications are contingent rather than guaranteed and demonstrated state capacity to penetrate systems previously assumed to be impervious.
Simultaneously, the pandemic generated a behavioural data yield of extraordinary richness. Remote work didn't just enable productivity monitoring — it captured the texture of domestic life, the patterns of attention, the social graphs of professional and personal relationships. This data was not incidental to the crisis response. It was, in effect, one of its durable outputs.
III. From Extraction to Replacement
The extraction of human data was never an end in itself. It was training material. The large language models and generative AI systems now being deployed were built on the captured expressions, creations, and interactions of billions of people — scraped, ingested, and laundered into statistical patterns that can now simulate human output without human involvement.
This is the endpoint of the enclosure. The commons of human creativity was first harvested without consent, then used to build systems that render the original creators economically redundant. Artists, writers, musicians, coders, designers — the very people whose work trained the models — now find themselves competing with infinite, near-free synthetic alternatives derived from their own labour.
The implications extend beyond the creative industries. As automation advances, the question becomes stark: in a world where machines can perform most economically productive tasks, what is the basis for human participation in the economy? What is the basis for human claims on resources, political voice, or social standing?
The trajectory, if uncontested, points toward a new feudalism. A small class of platform owners and capital holders. A managerial stratum that maintains the systems. And a vast population whose labour is unnecessary, whose data has been extracted, and whose claims on resources depend entirely on the sufferance of those above them. This is not a prediction of intent. It is a description of the structural tendency that emerges when labour becomes decoupled from economic necessity.
IV. Governance by Neglect
"When labour is no longer required, rights lose their material anchor."
This is the sentence that the political economy of automation forces us to confront. The Enlightenment secured rights partly through moral argument, but also through the leverage that labour scarcity provided. Workers were necessary. Their cooperation had to be purchased. Their discontent carried material costs. From this structural position, rights were won — not given.
What happens when that leverage evaporates? When a population's labour is no longer required for production, when their attention has already been harvested for training data, when their consumption is not necessary to sustain demand in an economy increasingly oriented toward luxury goods and positional assets?
The answer need not involve overt malice. It requires only the withdrawal of investment in systems that sustain populations who no longer generate returns. Healthcare systems that manage decline rather than restore function. Housing policies that permit extraction without construction. Education reduced to credentialing rather than capability-building. Social safety nets designed to maintain minimum subsistence rather than enable participation.
This is not conspiracy, it is Malthusian governance by neglect: the recognition that under conditions of automation and capital concentration, there are no structural incentives to maintain large populations at high standards of living. The outcomes — declining fertility, increased mortality in lower socioeconomic strata, concentration of longevity and capability in capital-rich populations — emerge not from policy intent but from policy absence.
Climate change compounds these dynamics. The coming decades will bring resource scarcity, mass displacement, and infrastructure stress. Adaptation will be expensive. Those with capital will purchase resilience. Those without will face systems designed for managed decline rather than recovery. The sorting is already underway.
V. Proof of Humanity
If the enclosure is of the human mind, then resistance must begin with reasserting what is irreducibly human — and building systems that make that assertion verifiable, legally meaningful, and economically consequential.
This is the concept of Proof of Humanity: not as a blockchain protocol, but as an existential credential. The capacity to demonstrate that a creative work, a decision, an expression, emerged from a human mind rather than a statistical model. In a world of infinite synthetic content, this becomes the scarce resource. In a world of automated labour, it becomes the basis for continued economic participation.
Any infrastructure for Proof of Humanity must address three distinct layers. These are not problems this thesis claims to solve. They are questions that must be answered — by technologists, legal scholars, policymakers, and the creative communities whose standing depends on the answers.
The Ontological Layer. What does it mean for something to be human originated? This is not a trivial question. Human creators use tools, including AI tools. They collaborate, iterate, and build on prior work. The ontological question concerns where the line is drawn between human authorship and machine generation — and whether that line can be defined in ways that are stable, meaningful, and resistant to gaming. The answer will likely involve gradations rather than binaries: degrees of human involvement, with different thresholds for different purposes. The law already handles analogous boundary problems — joint authorship, work-for-hire, derivative works — but whether these frameworks can be extended or whether new ones are needed remains contested.
The Epistemic Layer. How can human origin be demonstrated credibly? This is the technical challenge. Multiple approaches may satisfy it: cryptographic watermarking that embeds provenance signals into creative works at the point of creation; registry systems that establish chains of custody and attribution; attestation mechanisms that link identity to output; detection systems that distinguish human-originated content from synthetic content. Each approach carries trade-offs. Watermarking faces adversarial pressure. Registries raise questions of governance and access. Detection systems produce false positives and false negatives. The question is not which approach is perfect — none will be — but which can be made robust enough to support legal and commercial reliance, and at what cost.
This is an arms race, not a solved problem. The goal is not to guarantee provenance absolutely, but to raise the cost of circumvention sufficiently that compliance becomes the path of least resistance — and that non-compliance carries meaningful risk.
Authentication has never required perfection. Passports are forged; signatures are faked, financial systems are defrauded daily — yet we do not conclude that identity verification is pointless. We accept that authentication reduces fraud to levels compatible with functional systems. The alternative to imperfect authentication is not perfect authentication; it is no authentication at all — and therefore no standing, no attribution, no basis for legal or economic claims.
The same logic applies to human provenance. The question is not whether watermarking or registry systems can be made impervious to attack. They cannot. The question is whether they can raise the cost of circumvention sufficiently that compliance becomes the default, and whether they can provide evidentiary foundations for legal and economic enforcement. The infrastructure of citizenship and financial transaction operates on this basis. The infrastructure of human authorship must do the same — or authorship itself becomes unverifiable, and therefore uncompensated, and therefore economically extinct.
The Jurisdictional Layer. Under what legal regimes does demonstrated human origin have force? This is where Proof of Humanity meets enforcement. The EU AI Act establishes obligations around training data transparency and risk classification. Copyright regimes are being tested by litigation over unauthorised training. Data protection frameworks may provide grounds for action against behavioural capture. The jurisdictional challenge is to translate technical provenance into legal standing — to make synthetic impersonation and uncredited training actionable, and to embed compliance with human provenance standards into regulatory requirements rather than leaving them as market choices.
Regulatory arbitrage is a real risk. If major jurisdictions outside the EU do not adopt similar standards, the global internet may fragment, or extractive systems may simply relocate to deregulated zones. Partial coverage, however, still creates market segmentation that gives human-verified content a premium within regulated jurisdictions — and regulatory frameworks have a history of extraterritorial effect when platforms cannot easily serve different content to different regions.
The Capture Risk. Any system of verification carries the risk of becoming another form of enclosure — a luxury credential available only to those who can afford verification, creating a two-tier system where certified human creation commands premium prices while everyone else competes with machines. This is not a hypothetical concern; it is a central design constraint that any implementation must satisfy. To prevent this capture, the infrastructure must be open, interoperable, and accessible. It must be foundational rather than proprietary. It must be designed as civic infrastructure, not as a product. The specific mechanisms for achieving this — governance models, funding structures, technical architectures — are not prescribed here. They must emerge from collective deliberation among those who will build and use these systems.
VI. The Enlightenment Stakes
The Enlightenment's core claim was that humanity itself confers rights. Not birth, not land, not capital — but the bare fact of being human. From this claim flowed political equality, universal suffrage, human rights frameworks, the welfare state, public education, and the presumption that every person has standing to participate in collective decision-making.
These gains were not inevitable. They were fought for, over centuries, against entrenched power. And they were secured, in part, because human labour was necessary. The feedback loop between economic participation and political power gave ordinary people leverage. Strikes, boycotts, and collective action worked because production required cooperation.
Automation threatens to break this feedback loop. If production no longer requires human labour at scale, the material basis for political leverage dissolves. Rights become grants rather than claims — dependent on the continued willingness of capital holders to sustain systems that distribute resources to those who no longer contribute to production.
This is not an argument against automation per se. It is an argument for building new infrastructure that preserves human standing independent of labour contribution. Proof of Humanity is one component of that infrastructure. Universal basic income is another. Democratic reform that decouples political voice from economic productivity is another. The point is that the Enlightenment settlement must be actively extended to meet new conditions — or it will be quietly withdrawn.
If surveillance systems can predict and pre-empt dissent, collective action becomes impossible. If synthetic content floods the information environment, shared reality dissolves. If the basis for economic participation becomes capital ownership and algorithmic leverage rather than human labour and creativity, the political power of ordinary people evaporates. These are not distant possibilities. They are trajectories already underway.
VII. The Work Ahead
To resist the enclosure of the human mind requires action on multiple fronts: technical, legal, economic, and cultural. What follows is not a blueprint but an orientation — a set of directions that must be refined through collective experimentation and contestation.
Technical. Build the infrastructure of human provenance. The specific mechanisms — whether cryptographic watermarking, provenance registries, attestation systems, or approaches not yet conceived — will emerge from experimentation. What matters is that the infrastructure be open and interoperable, resistant to capture by platform monopolies or proprietary lock-in. The technical layer does not solve the economic problem; it creates the addressing system that makes economic flows possible. It is necessary but not sufficient.
Legal. Exploit emerging regulatory frameworks. The EU AI Act's August 2026 enforcement deadline creates a window for establishing human provenance as a compliance standard. Copyright litigation is testing the boundaries of fair use in training contexts. Data protection regimes may provide grounds for action against behavioural capture. The objective is to translate technical provenance into legal standing — to make compliance with human provenance standards a regulatory requirement, not a market choice.
Economic. Create economic mechanisms that channel value to verified human creators. This is where the infrastructure of provenance becomes materially consequential. Possibilities include licensing regimes for training data, collective bargaining structures for creative workers, platform cooperatives that distribute rather than extract value, and new models of patronage and subscription that bypass extractive intermediaries. The goal is to ensure that human provenance carries economic weight — that verified human creation commands compensation, not merely recognition. The technical layer enables the legal layer enables the economic layer. Each is necessary; none alone is sufficient.
Cultural. Name what is happening. The power of the first enclosure lay partly in its diffusion — no single villain, no single moment, just a gradual fencing that seemed natural and inevitable. To resist the second enclosure, we must make it visible, speakable, contestable. We must articulate what is being lost and why it matters — not as nostalgia, but as structural analysis. The enclosure of the human mind is not a metaphor. It is a process. And processes can be interrupted.
But naming alone is insufficient. The writers' and actors' strikes of 2023 demonstrated that collective action remains possible even as labour's structural leverage declines. The window has not closed. Naming must be coupled with organising — building coalitions across creative industries, academic institutions, civil society, and the technical communities whose participation any solution requires.
VIII. Conclusion: The Name We Give It
The Enclosure of the Human Mind. This is the name for the process we are living through: the extraction of human creativity, attention, and agency for the training of systems that render human participation optional; the erosion of democratic accountability in favour of algorithmic governance and capital flight; the normalisation of surveillance and the structural pre-emption of dissent; the withdrawal of investment from systems that sustain populations whose labour is no longer required.
Naming it does not stop it. But naming it makes it visible. And visibility is the precondition for the construction of countervailing infrastructure.
In the first enclosure, the peasantry lost their commons but gained, eventually, industrial wages, political organisation, and the welfare state. The material necessity of their labour gave them leverage. The second enclosure offers no such consolation. If human creativity and labour become economically redundant, there is no next stage of capitalism that reabsorbs the displaced. There is only exclusion — unless alternatives are built.
Proof of Humanity is not a complete solution. It is a necessary component of a larger infrastructure that must also include economic redistribution, democratic reform, and legal frameworks that establish human standing independent of productive contribution. But it is the component that preserves agency — the capacity to create, to be recognised as the source of creation, and to participate in the economy on that basis.
The Enlightenment's promise was that humanity itself is sufficient grounds for dignity, participation, and rights. That promise is now contested — not by philosophers, but by systems and the capital that deploys them. The question is whether we build infrastructure that extends that promise to meet new conditions, or whether we allow it to be enclosed.
The window is narrowing. The work is urgent. And the name of the work is clear.
This is the work.
The full thesis carries scholarly references (Boyle, Zuboff, Polanyi, Thompson, Acemoglu & Restrepo, and others). It forms the basis of the forthcoming book The Enclosure of the Human Mind from Manchester University Press, and of the Manchester Lecture with John McGrath of Factory International. Related reading: Genotone — Proof of Human.