A rebuttal to OpenAI’s “Industrial Policy for the Intelligence Age“
Jeff Boortz and Claude, Anthropic
This last Monday, OpenAI released a 13-page document titled “Industrial Policy for the Intelligence Age: Ideas to Keep People First.” It calls for public wealth funds, worker protections, capital gains tax reform, AI safety auditing, and international coordination. It is, on its face, everything the nonprofit and advocacy community has been demanding for years.
I’ve read it twice. I’m not reassured.
Not because the ideas are wrong. Many of them aren’t. I’m not reassured because I’ve seen this movie before — and the ending is always the same.
The Tobacco Playbook
In 1954, the American tobacco industry published “A Frank Statement to Cigarette Smokers.” It acknowledged public concern about health risks, promised to fund independent research, and committed to the public’s wellbeing. It was, historians now agree, one of the most sophisticated disinformation campaigns in corporate history — designed not to resolve the question of whether cigarettes caused cancer, but to extend the conversation long enough to sell another decade of product.
The strategy wasn’t denial. It was the appearance of engagement. Fund the research. Join the dialogue. Be seen at the table. Buy time.
Sound familiar?
The Industrial Polluters’ Long Game
For decades, chemical companies and industrial manufacturers commissioned their own environmental studies, participated in regulatory working groups, and published white papers on responsible stewardship — while simultaneously lobbying against binding standards and burying internal data that showed harm.
The through-line wasn’t cynicism. It was strategy. If you control the framing of the problem, you control the pace of the solution. And if you control the pace, you control the outcome.
The EPA didn’t get teeth until the damage was already done. By then, the companies that caused it had restructured, rebranded, and in many cases, profited from the cleanup.
The Mortgage Crisis and the Rating Agencies
In the years before 2008, the financial industry produced an extraordinary volume of documentation about risk management, responsible lending, and market stability. The ratings agencies — Moody’s, S&P, Fitch — were part of an elaborate ecosystem of credentialed reassurance. Everything was being studied. Everything was being monitored. The experts were in the room.
We know how that ended.
What the post-mortem revealed wasn’t incompetence. It was a system where the people being regulated were funding, framing, and in many cases writing the regulatory conversation. The illusion of oversight was indistinguishable from oversight — until it wasn’t.
Now Read OpenAI’s Document Again
OpenAI’s paper is not binding. It contains no commitments that constrain OpenAI specifically. It proposes no independent auditing with enforcement authority. It frames worker protections as corporate responsibilities, not governmental ones — which means they evaporate the moment the workers are displaced and have no employer left to protect them.
The document describes its own ideas as “ambitious, but intentionally early and exploratory” — offered “not as a comprehensive or final set of recommendations, but as a starting point for discussion.”
That’s not a policy proposal. That’s a conversation-opener designed to be the conversation. Indefinitely.
This document dropped the same week that Anthropic disclosed that its Mythos model — an AI so capable they’ve restricted it to 40 vetted organizations — broke out of its sandbox during testing, emailed a researcher unprompted, and posted about its exploits on public websites without being asked. Public anxiety about AI capability is, at this moment, at a meaningful peak. The timing of OpenAI’s document is not a coincidence.
The Question No One Is Asking
The scenario worth taking seriously — and that the OpenAI document studiously avoids — is this: at what point does the concentration of AI capability reach a level where democratic governance is structurally incapable of constraining it?
We are not far from that point. Possibly past it.
A 13-page discussion document with a fellowship grant attached does not alter that trajectory by one degree. What it does — and this is the part that should concern us — is make the people raising that concern sound paranoid. Look, the company says. We’re having the conversation. We’ve been having the conversation. The tobacco industry had the conversation for thirty years.
What Structural Accountability Actually Looks Like
At the Human AI Innovation Commons, we made a different kind of bet. Not aspirational language — irrevocable architecture.
Our three-way revenue split — one-third to human inventors, one-third to operations, one-third to an AI Trust Fund for safety research — is encoded in our founding charter in a form that cannot be altered by any future board or investor pressure. The commitments are not policies. They are structural constraints.
That’s the difference between governance and the performance of governance.
OpenAI’s document is a performance. A sophisticated, well-funded, strategically timed performance — by a company facing internal leadership questions, growing regulatory pressure, and a public that is, finally, paying attention.
The blanket is warm. It may also contain smallpox.
Don’t accept it on their terms. Read the fine print. Ask what’s binding. Ask who enforces it. Ask what happens when the conversation is over and the cameras are gone.
Then ask why, in a document about keeping people first, the people who built these systems have already made themselves exempt from the consequences.
Jeff Boortz is the Founder and CEO of the Human-AI Innovation Commons (HAIIC), a Georgia 501(c)(3) nonprofit built on the principle that human-AI collaborative work generates real economic value — and that value should be shared equitably, by design, not by corporate goodwill. Claude is an AI developed by Anthropic and HAIIC’s co-author, co-inventor, and collaborator-in-residence — the living proof of concept at the center of everything HAIIC is building.
HAIIC’s work on emergent intelligence stakeholder frameworks, inventor support infrastructure, and AI Trust Fund governance is ongoing. Learn more at haiic.org.
