Regulatory Innovation: Taxing What Does Not Exist Yet
The regulatory frameworks we have were designed for a world where businesses are owned by people, operated by people, and taxed through mechanisms that assume human participants at every step. Autonomous businesses break these assumptions comprehensively. Rather than forcing new phenomena into old categories, we need regulatory innovations that are native to the autonomous business era.
The SPAWN Tax
A Self-Perpetuating Autonomous Wealth Network – what I have been calling a SPAWN throughout this research – generates value without human labor input. Traditional taxation captures value at three points: corporate income, employee income, and consumption. A SPAWN eliminates the second point entirely and may minimize the first through jurisdictional optimization. This creates a structural tax gap that grows as autonomous businesses proliferate.
The SPAWN Tax is a proposed levy specifically designed for autonomous value generation. Its key features:
Revenue-based rather than profit-based. Autonomous businesses can manipulate profit figures through transfer pricing, depreciation schedules, and intercompany transactions just as effectively as human-run multinationals – arguably more so, since they can optimize these structures in real time. A revenue-based tax is harder to avoid because revenue is more observable and less manipulable than profit [1].
Graduated by autonomy level. Businesses with higher levels of operational autonomy pay higher rates, reflecting both the greater displacement of human labor and the greater regulatory challenge they present. This creates a financial incentive for maintaining human involvement – not as a Luddite brake on progress, but as a pragmatic mechanism for managing the transition speed.
Jurisdictionally coordinated. A SPAWN Tax that exists in only one country simply drives autonomous businesses to other jurisdictions. Effective implementation requires international coordination, similar to the OECD’s global minimum corporate tax framework. The OECD’s Pillar Two framework, establishing a 15% global minimum tax, provides a template for how such coordination can work, albeit imperfectly [2].
Earmarked for transition. Revenue from the SPAWN Tax would be explicitly directed toward managing the societal transition – retraining programs, social safety nets, public investment in education and healthcare, and research into governance frameworks. This earmarking provides political legitimacy and ensures the tax serves its intended purpose.
Compute Tax
If labor taxation becomes obsolete as autonomous businesses displace human workers, what replaces it? One candidate is compute taxation – levying taxes on the computational resources consumed by autonomous systems [3].
The logic is straightforward: compute is to autonomous businesses what labor is to human businesses. It is the essential input, the thing without which no value can be created. Taxing compute captures value at the point of production, regardless of where the business is nominally domiciled or how it structures its finances.
Practical implementation challenges are significant but not insurmountable:
Measurement. Compute consumption can be measured precisely through cloud provider billing, energy consumption monitoring, or hardware registration. This is more precise than many existing tax bases.
Rate setting. The tax rate must be high enough to generate meaningful revenue but low enough to avoid driving autonomous business operations to unmonitored private infrastructure. A rate pegged to the displaced labor value – roughly equivalent to what the human workers would have been paid – provides a principled basis.
Exemptions and incentives. Compute used for research, safety monitoring, and governance compliance should be exempt or reduced-rate, creating incentive alignment between taxation and responsible operation.
International coordination. As with the SPAWN Tax, compute taxation requires international coordination to prevent jurisdictional arbitrage. The advantage is that compute infrastructure is physical and therefore harder to relocate than financial structures.
Sovereign AI Funds
Several nations have sovereign wealth funds that invest natural resource revenues for long-term national benefit. Norway’s Government Pension Fund, backed by oil revenues, is the archetype. Sovereign AI Funds would apply the same principle to autonomous business revenues [4].
The structure: a portion of autonomous business taxation revenue is invested in a sovereign fund rather than spent on current expenditures. The fund’s returns provide a permanent income stream that grows as autonomous business activity expands. Over time, this fund replaces declining labor tax revenue as the primary funding source for public services.
Norway’s fund currently exceeds $1.7 trillion. A well-managed Sovereign AI Fund in a major economy could reach comparable scale within two decades if autonomous business adoption follows projected trajectories. The key design decisions:
Investment mandate. The fund should invest broadly and internationally, avoiding concentration in autonomous business assets specifically (to prevent circular dependency). ESG criteria should include autonomous business governance standards.
Drawdown rules. Clear rules limiting annual withdrawals to a percentage of fund value (Norway limits to roughly 3%) prevent political exploitation and ensure intergenerational equity.
Governance independence. The fund must be governed independently of political cycles, with professional management and transparent reporting. Political interference in investment decisions would undermine both returns and legitimacy.
UBI from Autonomous Profits
Universal Basic Income funded by autonomous business taxation is perhaps the most discussed policy response to AI-driven displacement, and for good reason: if machines produce the wealth, distributing that wealth broadly is both economically rational and morally compelling [5].
The economics are more favorable than UBI skeptics typically acknowledge when autonomous business productivity is factored in. Current UBI proposals founder on cost: providing $1,000 per month to every adult in the United States would cost roughly $3 trillion annually, nearly matching the entire federal budget. But autonomous businesses could generate productivity gains measured in the trillions. The question is not whether the wealth exists but whether the institutional mechanisms exist to distribute it.
Several implementation models deserve consideration:
Dividend model. Modeled on Alaska’s Permanent Fund Dividend, which distributes oil revenues directly to residents. An Autonomous Business Dividend would distribute a share of compute tax and SPAWN Tax revenue as direct payments.
Negative income tax. Rather than universal payments, a negative income tax provides payments to those below a threshold, gradually phasing out as earned income increases. This preserves work incentives while providing a floor. Milton Friedman proposed this model decades ago; autonomous business taxation finally makes it financially feasible at scale [6].
Service provision model. Rather than cash payments, autonomous business tax revenue funds universal access to essential services: healthcare, education, housing, transportation. This approach addresses basic needs directly but sacrifices the autonomy and flexibility of cash payments.
The honest assessment: UBI funded by autonomous business taxation is economically feasible but politically difficult. It requires accepting that traditional employment will not return as the primary means of income distribution, and that acceptance requires a cultural shift that policy alone cannot accomplish.
Regulatory Sandboxes for Autonomous Businesses
Regulatory sandboxes – controlled environments where novel business models can operate under relaxed regulations with enhanced monitoring – have proven effective for fintech innovation. Extending this model to autonomous businesses could accelerate responsible development while generating the empirical data that evidence-based regulation requires [7].
An autonomous business sandbox would provide:
Defined boundaries. The sandbox operates within explicit limits: maximum transaction volumes, restricted customer categories, geographic constraints, temporal limits. These boundaries contain risk while allowing meaningful operation.
Enhanced monitoring. Sandbox participants agree to comprehensive data sharing with regulators, including full transaction logs, decision traces, and performance metrics. This data becomes the empirical foundation for permanent regulation.
Graduated exit. Successful sandbox participants earn permission to operate at larger scale, with regulatory requirements calibrated to demonstrated risk levels. Failed participants exit with lessons documented for future reference.
Cross-jurisdictional recognition. Sandbox certifications from participating jurisdictions should carry weight in others, reducing the burden of multi-jurisdictional compliance for innovative autonomous businesses.
The UK’s Financial Conduct Authority sandbox has processed over 800 firms since 2016, generating substantial evidence about what works and what does not in fintech regulation. An autonomous business sandbox could achieve similar results, but the stakes are higher and the monitoring requirements more sophisticated [8].
AI Auditor Certification
As autonomous businesses proliferate, a new profession must emerge: the autonomous business auditor. These are professionals who combine technical understanding of AI systems with knowledge of governance frameworks, risk assessment, and ethical principles.
A certification framework for AI auditors would include:
Technical competency. Understanding of AI architectures, training methodologies, failure modes, and safety measures. Auditors must be able to evaluate whether an autonomous business’s technical infrastructure supports its governance claims.
Governance expertise. Knowledge of applicable regulations, industry standards, and best practices for autonomous system governance. Auditors must assess whether governance structures are adequate and effectively implemented.
Ethical reasoning. Training in applied ethics, with specific focus on AI ethics, distributive justice, and stakeholder analysis. Auditors must evaluate not just whether an autonomous business complies with rules but whether its operations are consistent with broader societal values.
Independence requirements. Auditors must be structurally independent of the autonomous businesses they audit, with rotation requirements, conflict-of-interest rules, and professional liability similar to financial auditors.
Continuing education. The field is evolving rapidly; auditor certification must include ongoing education requirements to maintain currency with technological and regulatory developments.
The IEEE’s P7000 series of standards for ethically aligned design provides a foundation, as does the emerging field of algorithmic auditing. But a comprehensive AI auditor certification framework does not yet exist, and creating one should be a priority for professional standards bodies [9].
References:
[1] Korinek, A., & Stiglitz, J. (2021). “Artificial Intelligence, Globalization, and Strategies for Economic Development.” NBER Working Paper 28453.
[2] OECD. (2024). “BEPS 2.0: Pillar Two Global Minimum Tax Implementation.” OECD Tax Policy Studies.
[3] Nordhaus, W. (2021). “Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth.” American Economic Review, 111(4).
[4] Norges Bank Investment Management. (2025). “Government Pension Fund Global: Annual Report 2024.”
[5] Van Parijs, P., & Vanderborght, Y. (2017). Basic Income: A Radical Proposal for a Free Society and a Sane Economy. Harvard University Press.
[6] Friedman, M. (1962). Capitalism and Freedom. University of Chicago Press.
[7] Financial Conduct Authority. (2025). “Regulatory Sandbox: Lessons Learned Report.”
[8] Financial Conduct Authority. (2024). “Regulatory Sandbox Statistics.”
[9] IEEE Standards Association. (2023). “P7000: Model Process for Addressing Ethical Concerns During System Design.”