A New Social Contract for an AI-Enhanced Economy

Doctrine in one sentence

Societies should secure mass economic agency by redistributing the surplus created by AI-driven productivity through transparent levies, citizen-owned investment, and universal basic services, while preserving democratic legitimacy and fiscal credibility.

Executive summary

The Unit-Cost Dominance thesis warns that widespread AI substitution could weaken the historic wage-demand circuit that underpins mass agency. It does not predict universal unemployment, nor does it prescribe any single institutional fix. The policy package proposed here is deliberately agnostic about the precise share of AI in any workflow and instead builds on observable, administrable levers that already exist in the United Kingdom and the European Union.

Four complementary planks form the core of the response:

1. AI-output levy – a transparent tax on the marginal productivity of AI systems that exceed a defined utilisation threshold, collected through existing corporation tax and digital services tax mechanisms.
2. Citizen-owned investment pool – expansion of the UK National Wealth Fund (NWF) and creation of an EU-wide “Digital Sovereign Fund” to acquire equity in AI-intensive firms, with dividend streams earmarked for a universal citizen dividend.
3. Universal basic services (UBS) – a modest expansion of housing, energy and digital access benefits funded by the levy and the dividend, guaranteeing a minimum level of economic agency regardless of labour income.
4. Lifelong learning accounts – individually portable vouchers that finance reskilling, up-skilling and credentialing, administered through existing education and skills agencies.

These measures have clear fiscal bases, can be implemented within existing legal frameworks, and deliver ordinary public-policy value—progressive taxation, public investment, and social protection—whether AI-driven dislocation is modest or dramatic. The design anticipates capture, avoidance and off-shoring, and includes trigger thresholds, review points and bounded-regret safeguards.

The policy package

1. AI-output levy

• Objective – capture a portion of the surplus generated when AI performs a task at a lower unit cost than human labour.
• Base – firms that report AI utilisation above a calibrated threshold (e.g., 30 % of output measured through the UK AI Registry and the EU AI-use notification system). The levy is levied on the incremental profit attributable to AI, estimated using a standard cost-adjusted return-on-assets formula already used in corporate tax assessments.
• Rate – an initial 2 % of the AI-adjusted profit, rising to 5 % if AI utilisation exceeds 60 % of total output. The rate is capped to avoid disincentivising adoption.
• Administration – integrated with HM Revenue & Customs (HMRC) and the European Commission’s Taxation and Customs Union (TCU) via automated data feeds from AI licence registers, cloud-compute import records and digital services tax filings.
• Revenue earmarking – 50 % to the citizen-dividend pool, 30 % to UBS, 20 % to the Lifelong Learning Accounts.

2. Citizen-owned investment pool

• UK National Wealth Fund (NWF) expansion – allocate an additional £10 bn from the AI-output levy to the NWF, directed toward equity stakes in high-growth AI-intensive firms, particularly those listed on domestic exchanges. The NWF can also use convertible instruments to preserve fiscal flexibility.
• EU Digital Sovereign Fund (DSF) – a multi-annual €30 bn pool funded by the EU-level AI levy (see below), managed by the European Investment Bank (EIB) under a governance structure that includes civil-society representation and member-state oversight.
• Dividend mechanism – net cash dividends from the NWF and DSF are pooled and distributed as a quarterly “Citizen Dividend” of £75 per adult in the UK and €50 per adult in the EU, indexed to inflation. The dividend is universal, not means-tested, preserving the principle of mass agency.

3. Universal basic services (UBS)

• Housing – expand the existing “Help to Buy” and social-housing guarantees to provide a rent-subsidy credit of up to £200 per week for low-income households, funded partially by the levy.
• Energy – introduce a capped energy-poverty rebate that offsets the median household electricity bill by 30 %, financed through the same revenue stream.
• Digital access – guarantee nationwide broadband of at least 100 Mbps, with a “connectivity voucher” of £30 per month for households lacking coverage, sourced from the dividend pool.

4. Lifelong learning accounts (LLA)

• Structure – each resident receives a £2 000 lifelong learning account, portable across providers, administered through the UK’s Skills Funding Agency and the EU’s European Social Fund Plus (ESF+).
• Use – the account can fund accredited courses, micro-credentials, apprenticeships and recognised up-skilling pathways, including AI-safety and ethics certifications.
• Funding – the AI-output levy contributes 20 % of the LLA pool, with the remainder drawn from existing skills budgets, ensuring no net fiscal pressure on the OBR.

5. Data-portability and competition safeguards

• Mandate – strengthen the UK Digital Markets, Competition and Consumers Bill and the EU Digital Services Act to require “AI-output portability” for large platform providers, reducing lock-in rents.
• Enforcement – empower the Competition and Markets Authority (CMA) and the European Commission’s Directorate-General for Competition to impose corrective duties and levy “platform-rent” fees on firms that impede data mobility.

6. AI safety verification authority (optional)

A modest, independent body (the “AI Safety Oversight Board”) would certify high-risk AI systems under the EU AI Act, collecting a small administrative fee (0.1 % of AI-output levy revenue) to fund audits and ensure the legitimacy of the AI-output measurement methodology. The board’s remit is narrow, limiting the risk of regulatory capture.

United Kingdom: first 24 months

Governance and oversight

• AI-Output Levy Committee – chaired by the Chancellor, includes representatives from Treasury, DSIT, CMA, trade unions and the Confederation of British Industry. Meets quarterly.
• Audit – National Audit Office (NAO) conducts annual audits of levy collections, fund allocations and dividend disbursements.

European Union and member states: first 24 months

Multi-level coordination

• EU-UK Joint Working Group – continues cooperation on AI registration, levy avoidance prevention and dividend harmonisation, respecting the post-Brexit trade framework.

Years 3 to 5 and dormant triggers

Doctrine in one sentence

Societies should secure mass economic agency by redistributing the surplus created by AI-driven productivity through transparent levies, citizen-owned investment, and universal basic services, while preserving democratic legitimacy and fiscal credibility.

Executive summary

The Unit-Cost Dominance thesis warns that widespread AI substitution could weaken the historic wage-demand circuit that underpins mass agency. It does not predict universal unemployment, nor does it prescribe any single institutional fix. The policy package proposed here is deliberately agnostic about the precise share of AI in any workflow and instead builds on observable, administrable levers that already exist in the United Kingdom and the European Union.

Four complementary planks form the core of the response:

1. AI-output levy – a transparent tax on the marginal productivity of AI systems that exceed a defined utilisation threshold, collected through existing corporation tax and digital services tax mechanisms.
2. Citizen-owned investment pool – expansion of the UK National Wealth Fund (NWF) and creation of an EU-wide “Digital Sovereign Fund” to acquire equity in AI-intensive firms, with dividend streams earmarked for a universal citizen dividend.
3. Universal basic services (UBS) – a modest expansion of housing, energy and digital access benefits funded by the levy and the dividend, guaranteeing a minimum level of economic agency regardless of labour income.
4. Lifelong learning accounts – individually portable vouchers that finance reskilling, up-skilling and credentialing, administered through existing education and skills agencies.

These measures have clear fiscal bases, can be implemented within existing legal frameworks, and deliver ordinary public-policy value—progressive taxation, public investment, and social protection—whether AI-driven dislocation is modest or dramatic. The design anticipates capture, avoidance and off-shoring, and includes trigger thresholds, review points and bounded-regret safeguards.

The policy package

1. AI-output levy

• Objective – capture a portion of the surplus generated when AI performs a task at a lower unit cost than human labour.
• Base – firms that report AI utilisation above a calibrated threshold (e.g., 30 % of output measured through the UK AI Registry and the EU AI-use notification system). The levy is levied on the incremental profit attributable to AI, estimated using a standard cost-adjusted return-on-assets formula already used in corporate tax assessments.
• Rate – an initial 2 % of the AI-adjusted profit, rising to 5 % if AI utilisation exceeds 60 % of total output. The rate is capped to avoid disincentivising adoption.
• Administration – integrated with HM Revenue & Customs (HMRC) and the European Commission’s Taxation and Customs Union (TCU) via automated data feeds from AI licence registers, cloud-compute import records and digital services tax filings.
• Revenue earmarking – 50 % to the citizen-dividend pool, 30 % to UBS, 20 % to the Lifelong Learning Accounts.

2. Citizen-owned investment pool

• UK National Wealth Fund (NWF) expansion – allocate an additional £10 bn from the AI-output levy to the NWF, directed toward equity stakes in high-growth AI-intensive firms, particularly those listed on domestic exchanges. The NWF can also use convertible instruments to preserve fiscal flexibility.
• EU Digital Sovereign Fund (DSF) – a multi-annual €30 bn pool funded by the EU-level AI levy (see below), managed by the European Investment Bank (EIB) under a governance structure that includes civil-society representation and member-state oversight.
• Dividend mechanism – net cash dividends from the NWF and DSF are pooled and distributed as a quarterly “Citizen Dividend” of £75 per adult in the UK and €50 per adult in the EU, indexed to inflation. The dividend is universal, not means-tested, preserving the principle of mass agency.

3. Universal basic services (UBS)

• Housing – expand the existing “Help to Buy” and social-housing guarantees to provide a rent-subsidy credit of up to £200 per week for low-income households, funded partially by the levy.
• Energy – introduce a capped energy-poverty rebate that offsets the median household electricity bill by 30 %, financed through the same revenue stream.
• Digital access – guarantee nationwide broadband of at least 100 Mbps, with a “connectivity voucher” of £30 per month for households lacking coverage, sourced from the dividend pool.

4. Lifelong learning accounts (LLA)

• Structure – each resident receives a £2 000 lifelong learning account, portable across providers, administered through the UK’s Skills Funding Agency and the EU’s European Social Fund Plus (ESF+).
• Use – the account can fund accredited courses, micro-credentials, apprenticeships and recognised up-skilling pathways, including AI-safety and ethics certifications.
• Funding – the AI-output levy contributes 20 % of the LLA pool, with the remainder drawn from existing skills budgets, ensuring no net fiscal pressure on the OBR.

5. Data-portability and competition safeguards

• Mandate – strengthen the UK Digital Markets, Competition and Consumers Bill and the EU Digital Services Act to require “AI-output portability” for large platform providers, reducing lock-in rents.
• Enforcement – empower the Competition and Markets Authority (CMA) and the European Commission’s Directorate-General for Competition to impose corrective duties and levy “platform-rent” fees on firms that impede data mobility.

6. AI safety verification authority (optional)

A modest, independent body (the “AI Safety Oversight Board”) would certify high-risk AI systems under the EU AI Act, collecting a small administrative fee (0.1 % of AI-output levy revenue) to fund audits and ensure the legitimacy of the AI-output measurement methodology. The board’s remit is narrow, limiting the risk of regulatory capture.

United Kingdom: first 24 months

Governance and oversight

• AI-Output Levy Committee – chaired by the Chancellor, includes representatives from Treasury, DSIT, CMA, trade unions and the Confederation of British Industry. Meets quarterly.
• Audit – National Audit Office (NAO) conducts annual audits of levy collections, fund allocations and dividend disbursements.

European Union and member states: first 24 months

Multi-level coordination

• EU-UK Joint Working Group – continues cooperation on AI registration, levy avoidance prevention and dividend harmonisation, respecting the post-Brexit trade framework.

Years 3 to 5 and dormant triggers

Funding and fiscal arithmetic

Missing arithmetic – precise AI-adjusted profit figures will require detailed modelling by HMRC and the European Commission; the above uses published profit aggregates and a plausible AI-use weighting.

Political coalition and public case

• Coalition partners – Labour Party (UK), Liberal Democrats, Green Party, trade unions (TUC), tech-friendly centrists, European Socialists, progressive business groups (e.g., Confederation of British Industry’s “Future-of-Work” taskforce).
• Public narrative – “A fair share of the AI boom for every citizen.” Emphasise that the levy does not penalise productivity, but ensures that the gains from AI are not captured solely by owners and shareholders.
• Losers – large AI-intensive firms facing a modest levy, high-profit shareholders, and short-term capital-gain focused investors.
• Compensation – dividend and UBS directly offset any net tax increase; lifelong learning accounts assure workers that up-skilling pathways are funded.
• Sequencing – start with levy and dividend to demonstrate immediate benefit; follow with UBS expansion once revenue streams stabilise; introduce anti-concentration surcharge later if market concentration emerges.

Durability and anti-capture design

1. Transparent base – AI-output is measured through statutory filing, cross-checked with customs data on imported compute and with cloud-service provider invoices, limiting avoidance.
2. International coordination – UK-EU Joint Working Group and OECD-style multilateral reporting standards reduce offshore shifting.
3. Independent oversight – NAO (UK) and European Court of Auditors (EU) audit levy collection and fund use annually; findings are published in full.
4. Sunset and review clauses – each plank includes a trigger-based review (e.g., revenue thresholds, concentration metrics) that can scale back or tighten measures.
5. Governance diversity – citizen representatives sit on the NWF and DSF boards, diluting capture by corporate interests.
6. Legal robustness – the levy is framed as a “profit-adjustment tax” rather than a “AI tax”, fitting within existing corporate tax legislation and avoiding the need for new primary legislation.

Legal and institutional obstacles

• UK – Need for Treasury Order to amend corporation tax regulations; potential parliamentary debate over “new tax”. The AI Registry requires statutory backing; could be introduced via secondary legislation under the Companies Act.
• EU – EU AI-output levy must be adopted as a Council Directive, requiring unanimity among member states—a known hurdle. A feasible workaround is to implement the levy through national legislation (e.g., Germany, France, Netherlands) and harmonise via the EU-wide “Common Reporting Framework”.
• State aid rules – NWF and DSF equity purchases must be notified under EU state-aid rules; however, the “risk-sharing” exemption for public-interest investments can be invoked.
• Data-privacy – AI-use declarations must respect GDPR; the registry can be designed as a “data-controller” with built-in privacy safeguards.

Failure modes, review and exit rules

All reviews are conducted by an independent “AI Economic Impact Panel” comprising economists, technologists, trade-unionists and civil-society experts. Their recommendations are submitted to Parliament (UK) and the European Parliament (EU) for legislative action.

Feasibility table

High, Medium and Low are assigned after weighing legal, political and administrative considerations as described.

What is genuinely new here

1. AI-output levy – unlike generic digital services taxes, this levy targets the productivity surplus generated specifically by AI, using an observable utilisation threshold rather than trying to measure “AI-share” of each workflow.
2. Citizen-owned investment pool – the combination of a sovereign wealth-style fund acquiring equity in AI firms and redistributing dividends directly to citizens is novel for both the UK and the EU, linking public capital markets to universal basic income.
3. Integrated “triad” of levy, dividend and UBS – the package ties fiscal, investment and social-provision elements together, ensuring that each component is funded and that the overall system remains beneficial even if AI impact is modest.
4. Trigger-based durability safeguards – explicit revenue-percentage thresholds and concentration metrics provide clear, falsifiable decision rules that can be acted upon without legislative overhaul.

Bottom line

The Unit-Cost Dominance thesis raises a credible contingency: AI could erode the wage-demand circuit that underpins mass economic agency. The policy package set out above offers a politically viable, fiscally credible and institutionally durable response that does not depend on precise measurement of AI’s share of labour, nor on universal restraint of AI adoption. By taxing the surplus of AI-driven productivity, investing the proceeds in citizen-owned equity, and redistributing the returns through a universal dividend and basic services, the UK and the EU can preserve democratic legitimacy, maintain state capacity and ensure that every household retains a genuine stake in the economy – whatever the pace of AI transformation.