Sam Altman’s observation that history’s guide suggests humanity will adapt to AI the way it adapted to the industrial revolution, finding new things to do and want while building ever more wonderful things for each other, drew some historical skepticism from Frank Diana, principal futurist at Tata Consultancy Services and writer at frankdiana.net, who joined Dan Proft on Chicago’s Morning Answer to offer a more nuanced reading of what technological revolutions actually look like and where AI fits in the broader pattern.
Diana said Altman’s framing skips over the Luddite movement of 1811 and the extended periods of human misery, displacement, and conflict that followed industrialization before the gains people now enjoy came to pass. Transitions between technological ages are not seamless. They are fits and starts, early adopters and late adopters, conflict and resolution, and the industrial revolution is a poor advertisement for a painless transition because it was not one.
His broader theoretical framework is that every age builds systems with limits that are invisible until pressures mount and constraints become visible. As those pressures accumulate, the existing system distresses, and historically one age has given way to another when its systems can no longer carry the load. He said we are in exactly that moment now, with pressures mounting across every domain simultaneously, in geopolitics, society, economics, education, and energy, in ways that are exposing the limits of the industrial-digital infrastructure we have built over the past century and a half. AI is a significant piece of the conversation, but it is one element among many producing the stresses.
His optimistic case for AI is that it could serve as what he called an intelligent orchestrator of systems that the current infrastructure cannot manage at scale. Energy is his clearest example: the strain that AI data centers are going to place on the electrical grid cannot be managed by the existing system, but AI itself could become the coordinating intelligence that optimizes energy distribution in ways humans and current grid management cannot. In education, personalized AI-driven learning could theoretically address declining aptitude levels and the mismatch between what current schooling produces and what a rapidly changing economy needs. He said every technological transition has produced both places where humanity is enhanced and places where it is diminished, and AI will do the same. The diminishing side in education, he said, is the misinformation and disinformation problem and the erosion of the capacity to evaluate the truth of what AI produces.
On Pope Leo XIV’s encyclical Magnificat Humanity, Diana said it mirrors what Pope Leo XIII did in 1891 with Rerum Novarum, which means new things, when the moral and ethical implications of industrialization required a framework that human institutions had not yet provided. He said what Leo XIV is doing is analogous, putting a moral and ethical framing on a technology that is not merely a tool but something with profound implications for current structures, institutions, and what it means to be human, while doing so well in advance of AI’s full impact rather than after the fact.
On extinction-level predictions, Diana was direct: he thinks they distract from the real near-term challenges AI presents and should not be the center of the conversation. He said there are plenty of legitimate challenges in what AI is doing and will do in the next several years that deserve serious attention, and that the consciousness and singularity discussions, while worth having in a limited way, pull focus away from those more immediate and tractable concerns.
On artificial general intelligence and recursive self-improvement, Diana said the recursive self-improvement concern is grounded in the valid observation that AI is already in a sense building itself as it learns, and that this creates genuine uncertainty about where it ends up. AGI as a concept refers to AI that can apply general intelligence across any domain rather than being narrowly useful for specific tasks, becoming smarter than humanity collectively rather than merely having superior recall and processing within defined areas. He said Marc Andreessen’s claim on the Joe Rogan podcast that AI already delivers answers across domains that exceed what any single expert can provide has some truth to it, given that AI has access to more information than any human could process. But he said his own experience using AI extensively is that removing the human from the loop produces errors, hallucinations, and unreliable outputs. True AGI, in his definition, is realized when you could let it operate without oversight and trust that the decisions and information it produces are sound. We are not there, and the distinction between having access to all of humanity’s accumulated knowledge and being able to build independently from that foundation is the line that separates current AI from the AGI that people actually mean when they invoke the term.