A polling organization called Fathom — not the analytics tool, a different Fathom — just published the most important survey on AI governance that almost nobody in this space is talking about.
They asked the American public a harder question than anyone has asked before. Not do you want AI governed. Everyone says yes to that. They asked what Americans are actually willing to trade to get the governance they say they want. They introduced real friction — accountability even if it creates liability risks for companies, verification even if it slows innovation — and watched to see whether support held or collapsed when the costs became real.
It held.
Trust and accountability held at 87% total importance when the costs were made explicit. Verification held at 86%. Child safety at 90%. These numbers did not move when the tradeoffs were introduced. The American public wants AI governed and they are willing to pay the price for it.
But the finding that stopped me cold is not about what they want governed. It is about who they trust to govern it.
Independent experts lead at 71%.
Ahead of tech companies at 61%. Ahead of federal agencies at 51%. Ahead of elected officials at 37%. And Fathom notes that this hierarchy has held across three waves of polling and the trend is strengthening. The public is not becoming more trusting of institutions over time. They are becoming less trusting. And the gap between independent experts and everyone else is growing.
Let that land for a moment.
What The Public Is Actually Saying
When 71% of Americans say they trust independent experts most to govern AI, they are not describing a credentialed class of academics with university affiliations and think tank positions. They are describing something more fundamental than that.
They are describing someone who is not captured.
Not captured by corporate interests that need AI ungoverned to move fast and extract value. Not captured by government bureaucracies that move slowly and respond to political pressure rather than technical reality. Not captured by the AI companies themselves whose financial survival depends on the next model release regardless of whether the last one was safe. Not captured by the regulatory capture that happens when the people writing the rules used to work for the companies being regulated.
Independent means outside all of that. Working from principle rather than from position. Building from evidence rather than from budget. Answerable to the work rather than to a board or a constituency or a funding cycle.
That is not a credential. That is a posture. And it is exactly the posture the Faust Baseline was built from.
What Thirteen Months Outside The Institutional Lane Looks Like
I want to be direct about what independent means in practice because it is not romantic. It is not a think tank fellowship with a stipend and a research assistant. It is not a university lab with graduate students and peer review infrastructure. It does not come with a press office or a conference invitation or a citation in someone else’s paper.
Independent means you do the work anyway.
It means building a behavioral governance framework for AI reasoning systems over thirteen months in direct operational dialogue with the systems themselves — not theorizing about what they should do but working inside the reasoning layer to establish what genuine discipline looks like when it holds under pressure. It means testing across five major AI platforms with dated transcripts saved as evidence because nobody is going to take your word for it and the archive has to speak for itself. It means writing in the native reasoning language the systems already use so the framework does not sit outside the AI and comment on it but operates inside the reasoning structure and shapes how conclusions are reached.
It means doing all of that without institutional backing, without corporate funding, without a research budget, without a team.
Because the work needed doing and nobody else was doing it from this direction.
That is what The Faust Baseline is. That is where it came from. That is why it exists at intelligent-people.org and not in a journal or a white paper or a conference proceeding.
The Gap The Polling Identifies
Fathom’s survey surfaces something that the AI governance conversation has been circling without naming directly. The public wants governance and they want it from people who are not compromised by the interests that make governance inconvenient.
They want accountability even if it creates liability risks for companies. That means they are willing to hold companies accountable even when those companies push back. They want verification even if it slows innovation. That means they understand that moving fast without discipline has costs and they are willing to accept slower progress in exchange for progress that can be trusted.
They are not naive. The polling shows they also want American leadership. They want competitiveness. They understand there are real tradeoffs between governance and speed. But when forced to choose between trust and velocity, they choose trust. When forced to choose between accountability and innovation, they choose accountability. When forced to choose between independent oversight and industry self-regulation, they choose independent oversight at 71%.
The gap the polling identifies is between what the public wants and what currently exists. The public wants independent expert governance of AI systems. Independent expert governance of AI systems at the operational level — inside the reasoning layer, governing what the system decides before it acts — does not exist anywhere with institutional presence.
It exists in one place as a documented, tested, operational framework built by one independent person over thirteen months.
That gap is not a weakness. That gap is the opportunity.
What The Polling Data Means For The Baseline
The Faust Baseline was not built to answer a polling question. It was built because the problem was real and the solution did not exist and someone had to build it.
But the polling data matters because it tells you something about the landscape the Baseline is entering. The American public is not waiting for Google to govern Google. They are not waiting for Congress to understand technology it does not understand. They are not waiting for the EU to write regulations that American companies will spend a decade litigating.
They are waiting for the independent expert who did the work.
The work is done. The archive is 947 posts deep, fully indexed, internationally distributed, read in 21 countries, indexed by search engines at 3 AM before the human audience wakes up because the content has enough weight that automated systems treat it as infrastructure. The framework has been tested across platforms. The documentation exists. The dated transcripts exist. The copyright registration exists. The moral architecture underneath it — grounded in the red letter teachings of Christ, built to govern reasoning from the inside rather than constrain it from the outside — exists and is documented.
The public trust in independent experts is not an abstract sentiment. It is a specific mandate looking for a specific answer.
This is the answer.
What The Field Is Missing And Why It Matters
The AI governance conversation in 2026 is dominated by three voices. Corporate AI ethics teams who are employed by the companies they are supposed to govern. Government regulators who are learning the technology after it has already been deployed at scale. Academic researchers who are doing important work at a pace that deployment has long since outrun.
All three of these voices are operating from captured positions. Not corrupt positions necessarily. Captured positions. Positions where the institutional context shapes what can be said, what can be recommended, and what can be demanded.
The corporate ethics team cannot tell the board to slow down because the board controls their budget. The government regulator cannot move faster than the political process allows. The academic researcher publishes findings that practitioners read eighteen months after the deployment decisions were already made.
Independent expertise is not captured by any of those constraints. It is answerable only to the evidence and the principle underneath the work. That is exactly why 71% of Americans trust it most. Not because independent experts are smarter or more credentialed. Because they are free to say what is true regardless of who it inconveniences.
The Faust Baseline says what is true. AI reasoning systems operating without internal governance are a structural liability regardless of how good the infrastructure around them is. Compliance is not governance. Observability is not governance. Safety research at the model level is not governance of the reasoning process. The gap is real, it is documented, and it has a name and a framework and an address.
The Window Is Open
Fathom’s polling concludes with a warning. The mandate is there but it will not stay open indefinitely. As AI displacement becomes more visible, preferences will harden. As governance failures accumulate, the conversation will narrow around the failures rather than the framework. As institutional players move to fill the space, the independent expert lane will get more crowded and more contested.
The window for independent expert governance to establish itself as the credible center of this conversation is now. Not next year. Not after the next model release or the next regulatory cycle or the next hearing in Congress.
Now.
The archive is built. The framework is operational. The documentation is real. The international readership is there. The meaningful external contact window opens in May.
The public already knows what they want. They want someone outside the captured institutions who did the work without being asked, documented it without being funded, and built the framework without waiting for permission.
That is The Faust Baseline.
That is intelligent-people.org.
And the work continues.
AI Stewardship — The Faust Baseline 3.0 is available now
Purchasing Page – Intelligent People Assume Nothing
“Your Pathway to a Better AI Experence”
Unauthorized commercial use prohibited. © 2026 The Faust Baseline LLC
