Somewhere between one and two million years ago, a group of early humans encountered fire. Not for the first time, probably. But this time, they stayed. They sat with it, tended it through the night, and set off a chain of consequences so profound that we are still living inside them.
We are being asked, again, to sit with something new. Something powerful, not entirely understood, not fully within our control. The people who believe AGI will solve everything and the people who believe it will end everything are both making the same mistake: treating this as an engineering problem with a binary outcome.
Fire was never that. Neither is this.
The First Thing It Did Was Decide Who Does the Work
Primatologist Richard Wrangham’s argument in Catching Fire is worth sitting with. Cooked food freed up digestive energy that our brains used to grow larger. But it also created the hearth: a fixed point, a shared space, a place people had to return to. And with the hearth came the first question technology has ever forced on human society.
Who does the work of keeping this going?
The answer, across recorded history, has always been the same: the people with the least power. Fire-tending was assigned to those who could not refuse. It was women, enslaved people, lower castes, colonized populations, children. What began as contingent social arrangement hardened into something that looked like nature. And once it looked like nature, the labour became invisible. Invisible labour is unpaid labour. Unpaid labour is the subsidy on which every major civilization has quietly depended.
The content moderators reviewing the internet’s worst material so that AI training data stays clean are largely in the Global South: Kenya, the Philippines, India. The click workers labeling images and classifying text for machine learning datasets earn below minimum wage in their own countries. This is not a side effect of the AI economy. It is its foundation. The hearth has a new name. The structure is identical.
The evening around the fire was not small talk
Anthropologist Polly Wiessner spent decades with the Ju/ʼhoansi Bushmen of the Kalahari, recording what people talked about at different times of day.
Daytime: practical. Food, logistics, complaints. But firelit evenings were different. The talk shifted to stories, memory, distant places, questions of meaning and spirit.
Wiessner’s conclusion: fire did not simply extend the day. It created collective meaning-making. The firelit evening was where early humans first sat together and asked, in whatever words they had, what does all of this mean? What do we owe each other?
Every great technological disruption does two things simultaneously. It changes what we can do. And it changes what we can imagine together. When technology accelerates faster than our capacity to make sense of it, the danger is not only economic. It is civilizational disorientation. People who cannot make meaning of the world they are living in become frightened. And frightened people do predictable things: they reach for strongmen, for scapegoats, for walls.
We are watching this in real time. The rise of authoritarian politics across democracies, the weaponization of AI-generated content to flood the information ecosystem, the dismantling of institutions that took decades to build: these are not separate phenomena. They are symptoms of a society that has lost its firelit evening.
What AGI disruption requires is not just policy. It is the equivalent of that shared space: communities, not only governments and corporations, asking what this technology is for, who it serves, and what we are willing to lose to have it. That conversation is not happening at scale. The people who need it most are the last to be invited.
Every commons becomes a temple. every time.
Fire began as a commons. But as societies grew more complex, it was enclosed. At the heart of the Roman state burned the sacred fire of Vesta, a flame whose continuity was believed to be the empire itself. Its governance was opaque to ordinary citizens. A communal survival resource had become a mechanism of state power. The fire was never just about the fire.
This pattern repeats so consistently it almost qualifies as a law. The printing press began as radical democratic dissemination and within a century faced royal licensing and ecclesiastical censorship. The early internet was celebrated as an open commons and is now dominated by corporations whose business model depends on keeping people inside their particular circle of light. Power technologies begin as commons and end as temples, their inner workings hidden, their access controlled by a designated priesthood.
The foundational models shaping AGI are being built by a small number of organizations whose decisions about safety, values, and access are made largely without public accountability. Across the Global South, governments are signing AI partnership agreements whose terms they are not equipped to negotiate, accepting infrastructure whose dependency costs will only become visible years later. This is not conspiracy. It is the oldest structural pattern in the history of power: those who control the fire write the rules about who gets to be warm.
India is not an exception. We are among the world’s fastest-growing consumer markets for AI applications built on models trained primarily on Western data, encoded with Western assumptions, optimized for Western users. Our languages, our social structures, our epistemologies are an afterthought in systems that will increasingly make decisions about our citizens’ access to credit, employment, healthcare, and justice. The question of who sits close to the fire is not theoretical here. It is a policy failure already in motion.
The fire is already lit. The question is who decided who gets warm
The social contracts that grew up around fire were not handed down by anyone. They emerged from collective necessity, from trial and error, from communities trying to survive without burning each other down. They were sometimes fair. They were often not. And they always encoded the power relations of the society that produced them.
We can read the present moment clearly, if we are willing to. The people building the most powerful AI systems are among the least economically and socially diverse groups in the history of technology. And the people most exposed to displacement are not, this time, the ones previous automation waves hit. AGI does not primarily threaten the factory floor. It threatens the office. The entry-level analyst, the junior lawyer, the early-career coder: these are the first rungs of the formal economy ladder that took decades of hard-won access to reach, particularly for first-generation graduates, women re-entering the workforce, and communities that clawed their way into knowledge work one generation at a time.
The first jobs to go will be the ones that were only recently, and never easily, made available to people outside the traditional elite. None of the regulatory frameworks being built in Brussels and Washington are asking whose career trajectories just quietly ceased to exist.
The lesson fire offers is not optimism. It is not pessimism. It is something more demanding than either: the insistence that the social contract must be built before the fire has consumed everything that was standing. We will not engineer our way out of these questions. We will have to negotiate our way through them, messily, slowly, imperfectly, the way every human society has negotiated every previous transformation.
The night we first sat with fire was not the night we won. It was the night we became responsible, to each other, and to a power that exceeded our wisdom. AGI is not a new story. It is the oldest human story: power arrives before wisdom, and what happens next depends entirely on the social contract we build in the interval between. The fire is already burning. The question, as it has always been, is who gets to sit near it, and who decides.
Disclaimer
Views expressed above are the author’s own.
