Lists of the most powerful people in AI tend to measure the same things: who has the models, who has the chips, who signs the biggest checks, and who can build the future. The same names keep recurring: the builders, the chipmakers, the financiers and, increasingly, the researchers shaping the technology itself. These ratings capture an important dimension of power. But it also raises a deeper question: As AI becomes more widely available, what forms of influence become more valuable, not less?
The pattern repeats often enough to seem like a law: When technology makes something abundant, power rarely disappears. And it moves to everything that becomes rare. The logic of power has always followed ownership of scarce inputs: land when kings had it, oil when barons had it, and computing power now that machines need it.
However, the irony is that even one of the most famous barons of all time was selling trust. John D. launched Rockefeller named his company Standard Oil because, as Ron Chernow has recorded, kerosene at the time was being so unevenly refined that it was killing thousands of people annually in lamp explosions. The name “Standard” itself was a promise: that the product could be trusted.
Economist Herbert Simon presented the modern version of the same idea in 1971: “A wealth of information leads to a poverty of attention.” The question worth asking about artificial intelligence is what makes it abundant, and what makes it scarce.
What makes AI abundant is intelligence, or at least a convincing imitation of it. The cost of operating the model at a fixed level of capacity It has fallen nearly 280-fold in nearly two years. Whatever the frontier laboratory ship was, the open weight equalizer had now arrived About four months later. Boundaries still matter. by Stanford Artificial Intelligence Indexthe best closed models continue to outperform the best open weight models. However, a lead measured in months is a lease, not a property.
What abundant intelligence produces is reasonable competence. Models can generate almost limitless possibilities. They cannot accept responsibility for choosing between those possibilities. Here judgment becomes rare. Public database now tracks More than 1,700 court rulings Involving fabricated citations Amnesty International. Even the statistics describing the flood have become part of the flood. One widely quoted claim is that 90 percent of online content will soon be machine-generated Trade book 2020. It has been footnoted in Europol report Then it was repeated as research for years. A 2024 revision removed the previous reference, but by then the claim had taken on a life of its own, and was cited online as part of a broader algorithmic drift.
The proof market is already under construction. Cloudflare, which accounts for nearly a fifth of the Internet, Changed the default in 2025 To prevent AI crawlers from ingesting human-generated pages unless they pay. Google is reportedly paying Reddit 60 million dollars annually For the right to be trained in human conversation. These are markets in origin, which is the naked truth of human origin.
But the source is only the beginning. Knowing that a human being has made something says nothing about whether it is right or safe to act on it. The deepest scarcity is judgment: the ability to discern, decide, and believe.
Television got there before analysts. In Game of Thrones, spymaster Varys poses a riddle. A king, a priest, and a wealthy man sit in a room with a shared “sword of sale,” and each orders him to kill the other two. Who lives and who dies? Varys answers: “Power lies where men believe it to be.” His answer points to an emerging group: the category of judgment.
The AI economy has created its own selling word: a mercenary capable of producing almost anything to anyone for a single symbolic price. The old mystery was what the seller thought; This person does not believe in anything and serves everyone, so the firm belief is transferred to those who act on its outcomes. The power lists rank the crown and the coin, but the puzzle says they were never the target.
A study conducted by KPMG and the University of Melbourne, which included 47 countries, found this Less than half of people trust artificial intelligenceWhile nearly two-thirds admitted to relying on its outputs without verifying them. This gap is important because trust acts like a form of capital. It accumulates slowly, loses quickly and does not return at the old price. Deloitte Refundable portion of a government contract In Australia after an AI-powered report was found to contain invented sources. Follow where trust is bought and sold, and a different set of gatekeepers will emerge.
The Big Four accounting firms are under construction Ensuring artificial intelligence in the paid service lineselling the right to say that the system has been tested. Insurance companies began at Lloyd’s of London Writing policies against AI hallucinationswith models evaluated before coverage is issued and payments triggered if agreed-upon performance thresholds are not met. Lloyd’s priced seaworthiness long before governments regulated it. Now it helps to price acceptable machine error.
Universities are Return to handwritten and proctored exams Because the article he took home no longer proved what he had done before. Sweden is Spending more than $100 million To return printed textbooks to its schools, after concerns that screen-first classrooms came at the expense of interest and deep reading, Norway went further, announcing… A near ban on artificial intelligence in primary schools. Different organizations respond in different ways, but all fall back on systems they can trust.
The obvious objection is that model holders are not transformed into commodities at all. They have become more powerful. The super scale plans approx $725 billion in capital spending The year is 2026, and companies are building frontier models It remains concentrated in the hands of a relatively few. This just sharpens the point.
A handful of companies have the power of permanent infrastructure. The race for the crown is crowded, international and largely invisible. But for everyone at the bottom of them, that is, almost every company, board and profession, models have become accessible inputs, priced in tokens.
This helps explain why adoption is so far ahead of readiness, even in organizations that don’t have the resources to deploy the technology well. In many executive teams, the motivation is not an attempt to dominate. It is the fear of losing power: of being the last to own assets that are no longer scarce.
Most lists of power naturally focus on the people who build and fund AI, but there is another form of influence emerging alongside them: the people and institutions whose judgment determines whether AI outputs are reliable and acted upon.
The judging class doesn’t take photos well. They include the proofreader whose signature still carries weight, the editor whose byline serves as collateral, the judge, examiner, underwriter, and scientist whose work survives scrutiny, and the executive whose signature no one feels compelled to double-check.
What they do have is the one asset that makes AI more valuable as it becomes more capable, because every improvement in a machine’s fluency raises the price of knowing when to believe it.
The test facing anyone in power today is simple. Ask what this situation is actually based on. If the answer is access, information, capital, or computing, it depends on something that technology makes increasingly easier to obtain. If the answer is that people act on your judgment without feeling the need to check it, then you have something much rarer.
The next chapter in the story of AI power may belong to the governance category. Most of its members would never think to include themselves.
Rahim Hirji is a future of work strategist, founder of SuperSkills Intelligence Company and author of Super Skills: The Seven Human Skills for the Age of Artificial Intelligence, Published by Kogan’s page and now.
