By — Ihina Yadav
Abstract
11th March 2026 marks the day when OpenAI CEO Sam Altman casually declared that he envisions “a future whereintelligence is a utility, like electricity or water, and people buy it from us on a meter.” This sparked great controversy and led to a chain of discourse where netizens noticed the dystopian future that awaited them, whereas Altman called it a business model. This article interacts with the idea of “metered intelligence” being a political-technological development with implications ranging across who has access to thinking, paying and opting out.
The Room Where It Happened
Before diagnosing what Altman intended to disclose from his now-viral statement, the most overlooked detail iswhere he disclosed it. It was at BlackRock’s U.S. Infrastructure Summit held in Washington, D.C., while conversing with the chairman of BlackRock’s Global Infrastructure Partners, Bayo Ogunlesi. The statement was made before a private audience of infrastructure investors, who include those who fund the water, power and telecommunication systems.
Altman was speaking about us, rather than speaking to us. This distinction is what forms the core of his argument. The CEO of the most dominant AI company in the global market presented the future of intelligence as a metered commodity in front of investors. The statement was not a mere prediction, but a pitch from his end. This is the beginning of building the infrastructure of thought, nd currently,it is being built for the shareholders first. We, as consumers, are simply on the demand side of the equation.
“On a Meter” is Not a Metaphor
According to Altman’s logic, it is evident that he thinks the metered future as democratising. It almost seems like development towards providing access, since by keeping the AI supply constrained, the prices will increase, making intelligence flow towards “rich people” only, while flooding the market makes it abundant and cheap. This sounds legitimate until we look closer and realise this is the oldest capitalism sales pitch: “we will sell you the thing we made unavoidable”.
The meter has already started running and has been running for some time. The price for a month of ChatGPT Plus access costs $20. The statement at BlackRock only describes an extension of this model wherein API is billed per token used, thus creating a system made as structural and invisible as the electricity bill.
What everyone is forgetting is that AI is nothing like water or electricity. A 24-hour supply of running water in your home does not think for you. It cannot write that mail to your boss, plan that itinerary for you, generate your arguments, or filter your news. The act of cognition, of reasoning, commodifying into a utility to be rented by a company, is crossing a threshold that no utility in history has ever crossed. The popular response to Altman’s vision was met by a simple argument: “AI is not electricity. Electricity powers the machine. AI will decide what the machinesees, remembers, filters, writes, recommends, negotiates and hides,”. Buying water makes you hydrated. Buying intelligence marks your cognitive decline.
You are the product.
If the product is free, you are the product. This old Silicon Valley maxim was described as surveillance capitalism by Shoshana Zuboff, which refers to the extraction of human experience as raw material to be used as behavioural training data for prediction markets. This concept fits perfectly in the case of AI, with a slight difference beingAI’s surveillance expands to learning from humans and charging humans for the outcome.
Reinforcement learning from human feedback is a framework that uses human judgment to train AI models. This means that every prompt sent, every correction made, and every disagreement expressed as feedback is a form of unpaid labour improving the product. At first, it seems as harmless as the autocorrect on the phone becoming slightlybetter, but eventually, one stands face to face with a meter that they helped build.
A post on the platform “X” called out the foundational theft committed during the course of training AI models in a single sentence: “We stole all your knowledge and art, and now we’re gonna put a meter on it and sell it back to you.You’re welcome.” Countless artists over the years have gotten heir individual art styles replicated and profited from millions of times due to models being fed their work without consent or compensation. In a nutshell, the creatives of our society unknowingly trained AI models that are designed to replace them, and in the future, also bill them.
The Dealer Doesn’t Use
The unsettling part is that the people selling this utility do not indulge in it themselves. The Waldorf School of the Peninsula in Los Altos is a private school that is known to ban screens entirely and house the “digerati” offspring.Furthermore, in the past, Steve Jobs shared how his children did not use the iPad either. These are not coincidences. Nellie Bowles explains this class implication by saying, “It could happen that the children of poorer and middle-class parents will be raised by screens, while children of Silicon Valley’s elite will be going back to wooden toys and the luxury of human interaction.”Don’t get high on your own supply, is a rule obliged by drug dealers. Apparently, so do AI CEOs. Protecting their children from a future of metered intelligence makes a statement far more striking than what Altman said at the summit. The architects of this phenomenon understand the dangerous implications of outsourcing thinking. It appears as though they are building a world within a world, the meter being for everyone else.
The hypocrisy here is class strategy. Utility is for consumption by us, and the opt-out is for those who built it.
Infrastructure Is Never Neutral
From a more philosophical perspective, Altman’s utility framing assumes infrastructure to be neutral. It treated intelligence similarly to how electricity simply flows, turning a blind eye to ownership, origin and other factors. The Jim Crow era also saw electricity being used as a medium to enforce segregation in the rural electrification planning during the 1930s in North Carolina. Highway systems were employed to destroy Black neighbourhoods that connected to White suburbs. nfrastructure cannot transcend its political nature.
Presently, AI infrastructure is being developed with the biases and assumptions that are believed by the people who build it. Datasets are trained to over-represent Western-male-centric perspectives, and labour is performed disproportionately by workers of the Global South on poverty wages. The International Scientific Report on AI safety discloses how this Western dominance immensely impacts the safety and reliability of AI across various areas. Furthermore, a study on the unequal distribution of training data reveals how Arabic, being the fifth most spoken language globally, comprises less than 1% of AI training data. Nations risk becoming consumers instead of producers because AI systems essential for sectors such as healthcare, finance, and education are hosted and riced by foreign countries.Therefore, across South Asia, Africa and MENA, the model of metered intelligence is equivalent to paying for domestic capability. The phenomenon of structural dependency existing in cloud computing is being mirrored and applied to thought itself.
Institutions that shape each successive generation are meant to reproduce ideology. Schools, language, law, media, etc,all act in accordance with a structural process that passes on social and cultural values. Artificial intelligence is the newest development in this area and functions similarly, the only difference being the sheer scalability and invisibility of its progression.
In another rather dystopian statement, Altman stated that “No kid born today will ever be more intelligent than artificial intelligence”, which, if it becomes the ground reality, will have some serious implications. What we teach AI today, AI will teach our children tomorrow. Every response generated carries and reproduces the thoughts, biases and assumptions we feed it today. An institution trained on hierarchical Western-centric data will present the worldview through a singular lens, thus making it the default, simply how things operate.
This is how ideology has always behaved. It normalises a belief system which affects and shapes how societies perceive their surroundings. In the near future, when, as Altman believes, intelligence will be commodified and function as a utility, these ingrained assumptions will transform from mere glitches to inherent features, becoming structural and self-reinforcing, only understood as common sense.
Conclusion: Citizens or Consumers?
This article does not pose the question of whether AI can or will become infrastructure. It already has. It does, however, urge one to ask oneself, what kind of relationship do we wish to have with AI? Will that be decided by the people trying to co-exist with it, or those who pitched it to the investors first? When infrastructure developed through biased ideology is bought and sold, it forms the foundation of a permanent bug in the architecture of inequality. It does not become neutral simply due to its ubiquitous nature; it promotes a bigger problem altogether. Artificial intelligence was not, and will never be, free. Unpaid labour, stolen creativity and human feedback combined and trained a modelthat is now charging us for what we once helped build. The men who devised this “business model” are aware of this. That is why their children go to schools with no screens. It is only wise for us to now delete our accounts as well.
About the Author:
Ihina Yadav is a third-year BA LLB student whose work sits at the intersection of law and emerging technology. She is particularly interested in AI regulation, tech and digital security, and dispute resolution, essentially, what happens when the legal system has to catch up with something it didn’t see coming.