Should Citizens Have a Vote on How AI Systems Are Used?

TL;DR: AI governance currently excludes the public despite AI's massive societal impact. Democratic participation through token-based governance models could restore citizen agency over systems that shape their lives.

Should Citizens Have a Vote on How AI Systems Are Used?

The most consequential technology of our time is being governed by people who were never elected to make decisions on our behalf. AI systems now influence hiring, lending, healthcare, criminal justice, and information access for billions of people. Yet the fundamental choices about how these systems work—what they optimize for, what data they use, what biases they embed—are made by a small group of technologists and executives operating behind closed doors.

This is not just a market failure. It’s a democratic failure. Citizens deserve a voice in decisions that shape their lives, especially when those decisions affect entire societies. The current system of AI governance—a technocracy split between corporate boardrooms and regulatory agencies—fundamentally excludes the people most affected by AI’s impact.

What Is AI Governance Today?

AI governance today operates as a technocracy where technical experts and corporate leaders make decisions about systems that affect billions without meaningful public input. Current governance structures center on internal company policies, industry self-regulation, and government agencies staffed by technical specialists rather than democratic processes.

The conventional wisdom holds that AI is too complex for public participation. Tech leaders argue that citizens lack the technical knowledge to make informed decisions about neural network architectures, training methodologies, or algorithmic parameters. Policymakers defer to technical experts, creating regulatory frameworks that prioritize industry input over public consultation.

This technocratic model assumes that technical optimization naturally serves public interests—that better accuracy, efficiency, or capabilities automatically translate to better outcomes for society. Under this view, the role of governance is to ensure AI systems work as intended, not to question whether their intended purposes align with public values.

Why Technocratic Governance Fails

The Values Problem

Technical optimization is never value-neutral. When Facebook’s algorithm prioritizes engagement, it embeds specific assumptions about what content people should see. When hiring algorithms screen resumes, they encode particular definitions of merit and qualification. When recommendation systems surface information, they shape what knowledge people access.

These are fundamentally political choices disguised as technical ones. The decision to optimize for user engagement over information quality, for example, reflects a value judgment that private profit matters more than public discourse. Citizens never consented to this trade-off.

Consider the recent case of AI-powered content moderation systems. Platforms like Twitter and Facebook use AI to automatically remove content deemed harmful. But “harmful” according to whom? The algorithms reflect the values and biases of their creators, not democratic consensus about appropriate speech boundaries.

The Representation Gap

Neither corporations nor governments adequately represent public interests in AI development. Corporate governance structures prioritize shareholder returns, not citizen welfare. A 2025 study by the Stanford Institute for Human-Centered AI found that 73% of AI product decisions at major tech companies involved no consultation with affected user communities.

Government regulation suffers from different but equally serious representation problems. Regulatory agencies like the FTC or NIST are staffed primarily by technical experts and lawyers, not representatives accountable to voters. The AI Safety Institute, established in 2024, includes no mechanism for citizen input despite regulating systems that affect everyday life.

This creates what political scientists call “democratic drift”—the gradual shift of political decisions away from democratic institutions toward technocratic ones. AI governance exhibits this pattern in extreme form.

The Scale Problem

AI systems operate at unprecedented scale, affecting millions of people simultaneously through automated decisions. Traditional governance mechanisms—whether corporate boards or regulatory agencies—lack the legitimacy to make choices with such broad social impact without democratic authorization.

When a single algorithm change affects how billions of people access information, that decision carries more political weight than most legislation. Yet it’s made by engineering teams following metrics set by executives, not through any process that resembles democratic governance.

The Case for Citizen Participation

Historical Precedent

Democratic societies have successfully included citizens in complex technical decisions before. We vote on nuclear power policies despite their technical complexity. We participate in environmental regulations through public comment processes and citizen review panels. We elect representatives who make decisions about military technology, healthcare systems, and financial regulations.

The argument that AI is “too complex” for democratic input reflects technocratic bias, not empirical evidence about citizen capacity. Research by the Deliberative Democracy Consortium shows that informed citizen panels consistently make reasonable judgments about technical policies when given appropriate information and deliberation time.

Values Alignment

Citizens are uniquely qualified to make value judgments about AI systems because they’re the ones who live with the consequences. Technical experts can optimize systems for specific metrics, but they can’t legitimately decide what metrics matter most to society.

The Cambridge Analytica scandal illustrated this principle clearly. Facebook’s data scientists could optimize for user engagement and advertiser value, but they couldn’t legitimately decide whether election manipulation represented an acceptable trade-off. Only democratic processes can make such fundamental choices about social priorities.

Practical Implementation

Modern technology enables new forms of democratic participation that weren’t previously feasible. Token-based governance systems allow for continuous, granular citizen input on specific AI system parameters. Platforms like Perspective AI demonstrate how users can directly vote on model selection, training data sources, and system behavior through POV token governance.

These systems solve traditional problems with direct democracy—like the difficulty of organizing mass participation or ensuring informed deliberation—through digital infrastructure that makes participation scalable and ongoing rather than episodic.

Addressing the Complexity Objection

Technical vs. Political Decisions

The strongest objection to citizen participation in AI governance is complexity: AI systems involve intricate technical details that most people don’t understand. This objection confuses technical implementation with political choice.

Citizens don’t need to understand transformer architectures to have legitimate preferences about whether AI systems should prioritize accuracy over fairness, or privacy over personalization. These are value judgments, not technical determinations. We don’t require voters to understand monetary policy mechanics to participate in elections that determine economic priorities.

A practical governance model would separate technical implementation from value specification. Citizens could vote on principles and priorities—what outcomes AI systems should optimize for, what trade-offs are acceptable, what uses should be prohibited. Technical experts would then implement systems that achieve those democratically-determined goals.

Informed Participation

Meaningful citizen participation requires information and deliberation, not just polling. Successful models like Ireland’s Citizens’ Assembly on abortion or France’s Convention Citoyenne pour le Climat show how ordinary people can engage with complex issues when provided expert input, structured deliberation, and adequate time.

AI governance could adopt similar approaches: citizen panels that hear from technical experts, affected communities, and various stakeholders before making recommendations about AI system design and deployment. Digital platforms could enable broader participation through structured online deliberation and voting systems.

This addresses legitimate concerns about uninformed participation while preserving democratic control over fundamental value choices. The goal isn’t to have citizens design neural networks, but to ensure AI systems serve democratically-determined public purposes.

Token-Based Governance Models

Beyond Traditional Democracy

Blockchain-based governance tokens offer a new model for citizen participation in AI systems. Unlike traditional democratic processes that operate through periodic elections and representative institutions, token-based governance enables continuous, direct participation in specific decisions.

Users of an AI system could hold governance tokens that give them voting rights over system parameters, model updates, and policy changes. This creates a direct relationship between AI use and AI governance—the people most affected by system decisions have the strongest voice in making them.

Perspective AI exemplifies this approach through its POV token system, where users can vote on model selection, data sources, and platform policies. This creates accountability between AI developers and users that doesn’t exist in traditional corporate or regulatory structures.

Distribution and Legitimacy

The key challenge in token-based governance is ensuring legitimate token distribution. If tokens concentrate among wealthy buyers or early adopters, the system reproduces rather than solves existing inequality problems.

Successful models distribute tokens based on usage, contribution, or democratic criteria rather than purchasing power alone. Some platforms reserve token allocations for affected communities, use quadratic voting to limit plutocratic control, or implement time-locked distribution systems that prevent rapid concentration.

The goal is creating governance systems where influence correlates with stake in the system’s outcomes rather than financial resources or technical expertise alone.

What This Means for AI’s Future

Competing Visions

The question of citizen participation in AI governance reflects a broader choice about AI’s development trajectory. The current path leads toward AI systems controlled by a small number of powerful actors—whether governments, corporations, or technical elites. The alternative path leads toward AI systems that are accountable to their users and society more broadly.

This choice will determine whether AI becomes a tool for democratic empowerment or technocratic control. Systems designed with citizen participation embed different assumptions about power, accountability, and purpose than systems designed through corporate or bureaucratic processes.

Decentralization as Democracy

Decentralized AI platforms represent one approach to democratizing AI governance. By removing single points of control and enabling user participation in system decisions, these platforms create structural conditions for more democratic AI development.

However, decentralization alone isn’t sufficient for democracy. Decentralized systems can still concentrate power among technical elites or wealthy token holders. True democratic governance requires intentional design choices that prioritize broad participation and accountable decision-making.

The most promising approaches combine decentralized infrastructure with explicit democratic governance mechanisms—token distribution systems that ensure broad participation, governance processes that include deliberation alongside voting, and accountability structures that connect system performance to user welfare.

The Path Forward

Immediate Steps

Citizens and democratic institutions don’t need to wait for permission to participate in AI governance. Existing tools—from regulatory comment processes to shareholder advocacy to consumer choice—create opportunities for democratic input on AI development.

More ambitiously, supporting platforms that experiment with democratic governance models creates proof-of-concept examples for broader adoption. When users choose AI systems that offer governance participation over those that don’t, they create market incentives for democratic design.

Institutional Innovation

Longer-term democratic AI governance requires new institutions designed for the digital age. This might include citizen panels with ongoing oversight over AI systems, democratic input processes for AI regulation, or hybrid governance models that combine representative and direct democratic elements.

The goal isn’t to slow AI development but to ensure it serves democratic purposes. Democratic governance may add deliberation time to AI deployment, but it prevents the much larger costs of building systems that society ultimately rejects or that exacerbate social problems.

The choice between technocratic and democratic AI governance is ultimately a choice about what kind of society we want to live in. Technology shapes society, but society can also shape technology—if we insist on our democratic rights to do so.

Democratic participation in AI governance isn’t just possible—it’s essential for ensuring AI systems serve human flourishing rather than concentrating power among technical and economic elites. The question isn’t whether citizens are smart enough to participate in AI governance. The question is whether AI systems are legitimate enough to operate without their consent.

FAQ