AI ethics experts have spent years debating abstract principles like fairness and transparency, but they may be solving the wrong problem. Instead of starting with expert-designed ethics codes, policymakers should ground AI governance in public values: the deeply held beliefs that nearly all citizens already share, even when they disagree on implementation details.

What Are Public Values, and Why Do They Matter for AI?

Public values are the common sense of a political community, according to policy scholar Barry Bozeman. They represent the goals that almost everyone endorses, even if they disagree on how to achieve them. These values are embedded in constitutions, major legislation, court decisions, and the cultural stories a society tells itself.

In a survey of more than 2,000 U.S. citizens, Bozeman found striking consensus on fundamental values. Over 90% of respondents supported substantive principles including freedom of speech, physical liberty, equal access to civil rights, freedom of religion, gender equality, and physical safety and security. Over 80% supported protection of minority interests, access to health care, economic opportunity, and privacy.

The challenge is that while everyone agrees on these broad values in theory, people disagree sharply about what they mean in practice. Everyone supports "fairness," but Americans have fought bitterly over what fairness means in hiring, college admissions, and reparations. Similarly, "promoting human values" sounds good until you realize that liberty and equality can conflict with each other.

Why Do Current AI Ethics Principles Fall Short?

The consensus around AI ethics principles has grown impressive in recent years. Scholars and industry leaders have agreed on a framework that includes privacy, accountability, safety and security, transparency and explainability, fairness and nondiscrimination, human control of technology, professional responsibility, and promotion of human values.

But these principles often remain abstract and disconnected from how people actually live. When the nonprofit Center for AI Safety asked 100 prominent computer scientists and AI company CEOs to agree on something concrete, they could only reach consensus on a single sentence: "Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war." The problem is clear: such broad statements don't actually guide how to build better, more trustworthy AI systems.

The real issue is that different people have different definitions of responsibility, ethics, and benefits. What benefits one group often harms another. Managers and workers have conflicting interests. Competing firms want different rules. Urban and rural residents face different challenges. Rich and poor communities experience technology differently. In these contexts, agreement about AI ethics can only be reached on the tamest of terms.

How Can Policymakers Better Engage Public Values?

Rather than relying solely on opinion polls, which capture only a snapshot of views on a rapidly changing issue, experts recommend using deliberative methods. These approaches give citizens space to learn about technologies, discuss them with their peers, and take time to articulate their own priorities and reasoning.

• Public Forums: Bring together diverse community members to discuss AI applications and their implications for society, allowing citizens to voice concerns and priorities directly.
• Scenario Planning Methods: Present citizens with realistic future scenarios involving AI to help them think through consequences and articulate what outcomes they want to avoid or pursue.
• Integrating Social Researchers: Include humanistic and social science researchers alongside engineers in technical development, ensuring that human values are considered from the start rather than added later.
• Deliberative Engagement: Create structured conversations where citizens can learn, discuss with peers, and develop considered positions that are less prone to rapid changes in opinion.

These deliberative methods help citizens and researchers discover priorities together and connect today's technological issues to more durable conceptions of the good society. Decades of research in technology assessment, responsible innovation, and anticipatory governance have shown that these approaches work better than top-down expert pronouncements.

What Real-World Agreement Already Exists on AI?

Even in a deeply polarized America, there is surprising consensus on some AI applications. According to Pew Research data cited in the source material, few Americans support using AI to advise people about faith in God, matchmaking, or governing the country. Congress has already acted on one area of clear public concern: the TAKE IT DOWN Act of 2025 prohibited the dissemination of both real and AI-generated nonconsensual explicit content of individuals, addressing a growing and highly disturbing form of harassment.

These examples show that when policymakers identify genuine public values and build policy around them, they can move forward even amid broader polarization. The potential for political realignment exists, but only by engaging with one another and with the situation at hand.

Why Should AI Ethics Be Debated in Communities, Not Just Boardrooms?

AI ethics principles can, and hopefully will, eventually reflect more than just the discourse of experts and computer scientists. But this will only happen by being debated, hammered out, implemented, and modified in actual communities, governments, and businesses.

Public values are not eternal, unchanging, or universal. But they tend to be widely recognized and change slowly, emerging as they do from the long sweep of political contestation, negotiation, and compromise. Decisionmakers can get a head start by building upon public values, which have already emerged from this process of iterative trial, implementation, and deliberation.

The implication is clear: rather than waiting for perfect consensus on abstract principles, policymakers should invest in mechanisms for discovering and engaging the public values that already exist in their communities. This approach is more productive than chasing the lowest common denominator or hoping that expert ethics codes will somehow translate into trustworthy AI systems. Public values provide the foundation; implementation and accountability follow.