Dario Amodei’s recent essay on AI policy is worth reading, even if you do not agree with every part of it. I found myself agreeing with most of it. The article covers a wide range of issues, including the rapid pace of AI development, the possibility of transformative economic growth, national security implications, infrastructure demands, workforce disruption, and the need for thoughtful public policy. At its core, Amodei argues that AI is advancing exponentially and that governments need to prepare for a future that may arrive much sooner than most people expect.
AI is advancing quickly. It is moving faster than most business owners realize, faster than many institutions can adapt, and quite possibly faster than governments can respond in any meaningful way. That part did not surprise me.
As I thought through Amodei’s arguments, I found myself less concerned with technical expertise and more concerned with trust. Industry understands AI but has incentives. Government has authority but struggles with speed and understanding. Citizens may lack technical knowledge, but they are the ones who will live with the consequences.
In this article, I explore:
- Why expertise alone is not enough to create public trust
- The challenges government faces in keeping pace with AI
- How industry incentives complicate AI governance
- Why ordinary citizens deserve a meaningful role in the conversation
- Whether AI governance is ultimately a trust problem rather than a technology problem
The part that caught my attention came when the discussion turned to governance. For the past few years, I have told audiences that one of the defining characteristics of AI is its pace of change. Every time I update a presentation, new capabilities appear, new tools emerge, and new concerns surface. The conversation shifts so quickly that examples from six months ago can feel outdated. That raises an uncomfortable question: if AI is moving this fast, who exactly is supposed to govern it?
Expertise Is Only Part of the Problem
A common answer is to place experts at the center of the process. On the surface, that makes sense. The people building frontier AI systems understand them better than almost anyone else, and any serious conversation about AI governance needs their knowledge in the room. Companies like Anthropic, OpenAI, Google, and others possess technical insight that policymakers and the public simply do not have.
The problem is that expertise does not automatically create trust. These companies are not outside observers. They are participants in the market they are helping shape. That does not mean they are acting in bad faith, but it does mean they carry incentives into the conversation. Every organization has interests to protect, markets to defend, and futures to position itself for.
As I read Amodei’s essay, I kept coming back to a simple concern. If industry plays a major role in governing AI, who governs the governors? The challenge is not only finding people who understand the technology. The challenge is creating a process that the public can believe is not captured by the same forces it is meant to oversee.
Government Faces a Different Challenge
Government enters the conversation with a different set of strengths and weaknesses. It has legal authority. It can create standards, enforce rules, and establish consequences when companies cross lines. That role matters, and I do not think AI governance works without it.
What government struggles with is speed. This is not a partisan observation. It is a structural one. Government was designed to deliberate. It studies problems, forms committees, holds hearings, gathers input, debates proposals, and revises language. Those processes exist for good reasons, but AI does not operate on that timeline. While policymakers study one generation of technology, researchers may already be working on the next.
There is another challenge as well. Most elected officials are not AI researchers, software engineers, or technology founders. To understand what is happening, they depend on advisors, experts, lobbyists, and industry representatives. That dependence creates another layer of influence. Once again, the question becomes: who influences the influencers?
The Missing Voice
Most discussions about AI governance seem to revolve around two groups: industry and government. Industry knows the technology. Government has authority. But a third group receives far less attention, even though they have the most at stake.
That group is ordinary citizens. They do not need to understand large language models or evaluate technical architecture to deserve a role in the conversation. They matter because they are the people who live with the consequences. AI will affect their – OUR – jobs, communities, schools, creative work, privacy, energy costs, and sense of purpose. Those are not abstract policy categories. They are the texture of daily life.
When ordinary people talk about AI, the conversation often sounds different from what you hear in policy circles or technology conferences. They worry about job loss. They worry about data centers consuming water and electricity. They worry about artists, writers, teachers, and small business owners watching machines absorb pieces of their work. Some people dismiss those concerns as fear, but fear is not always irrational. Often, fear points toward something people value.
A person who fears job displacement values economic security. A person who fears the loss of creative work values human expression. A person who fears surveillance values privacy and autonomy. Those concerns may not tell us how to build AI systems, but they can tell us what those systems should protect.
The Risk of Capture
The AI governance problem is often framed as a lack of expertise. I am not sure that is right. We have experts. In fact, we may have more experts than we know what to do with. The deeper problem is capture.
Industry can be captured by profit and competition. Government can be captured by lobbyists, donors, and political incentives. Advocacy groups can be captured by funders and institutional agendas. None of this requires cartoon villains. It only requires normal human systems doing what they tend to do over time. The deeper problem is capture.
That is why citizen participation matters. It introduces a different kind of pressure into the system. Citizens bring lived reality into a conversation that can otherwise become too technical, too political, or too financially motivated. They bring fear, yes, but they also bring memory, conscience, and a sense of what people do not want to lose.
A More Human Table
I do not think citizens should write technical standards or pretend to understand every detail of frontier AI development. That would be its own kind of theater. But I do think citizens need a meaningful seat at the table, perhaps in a form closer to a jury than a political body. Limited terms. No campaign funding. No professional class of AI governors. No easy path for money to turn public oversight into another influence machine.
A healthier model would include industry, government, and citizens, with each group balancing the others. Industry would bring technical knowledge. Government would bring authority and enforcement. Citizens would bring legitimacy and lived experience. No single group should dominate the process, because no single group can be trusted with the whole thing.
The same principle applies in business. I tell people that if they are going to use AI, they need to put their own heart and soul into it. The tool should reflect what they value, how they think, and the people they serve. AI without that human grounding becomes efficient, but empty.
America’s Soul Belongs in the Conversation
AI governance cannot only be about safety tests, economic projections, or national competitiveness. Those things matter, but they are not the whole story. AI will shape work, creativity, education, leadership, privacy, relationships, and the way people understand their own value. This makes this a human issue before it is a technical one.
If AI is going to help shape America’s future, then America’s heart and soul need to be present in the process. That means we the people need more than reassurance from experts, companies, and committees. We need a real voice in the systems that will shape how we work, create, learn, and live.
The harder question may not be how to regulate AI. The harder question may be how to build a form of governance people can actually trust.
Cinguaris, AI Solutions with Soul