Is artificial intelligence really artificial? According to Blaise Agüera y Arcas, Google’s CTO for Technology and Society, the answer might surprise you. Speaking at a Harvard Law School Berkman Klein Center event, Agüera y Arcas suggested that the evolution of AI systems and human intelligence are not just similar — they may, in fact, share the same computational foundations.
Agüera y Arcas, author of the new book “What Is Intelligence? Lessons from AI About Evolution, Computing, and Minds,” explained that brains — both human and artificial — evolve through increasing computational power. “If we rewind 500 million years,” he said, “we see only creatures with very small brains. Go back a billion years, and there are no brains at all.” His central claim: brains are computers, not metaphorically, but literally computational systems that process information, transform inputs, and generate predictions — the same functions performed by AI.
This provocative idea challenges traditional distinctions between biological and artificial cognition. As Agüera y Arcas put it, “The premise of computational neuroscience is not that brains are like computers — it’s that they are computers.” His framework builds upon the foundational theories of Alan Turing and John von Neumann, pioneers who explored self-replication and universal computation. He also draws inspiration from evolutionary biologist Lynn Margulis, whose symbiogenesis theory argues that cooperation — not competition — drives complex evolution.
Agüera y Arcas applies this theory to both biology and artificial intelligence, arguing that when two computational systems interact symbiotically, they create a parallel computing environment that grows exponentially in complexity — much like neural networks. “Life was computational from the start,” he explained. “It gets more complex over time through symbiogenesis, because when computers start cooperating, they perform massively parallel computations — just like neurons in a brain.”
At Google, Agüera y Arcas conducted experiments that modeled this process in silico. Using a programming language with only eight basic instructions, he simulated millions of random interactions. Over time, self-replicating and self-improving digital entities emerged — demonstrating how complex systems can evolve from simple beginnings, much like the origin of life itself.
Beyond the technical, Agüera y Arcas connects intelligence to social cooperation. Drawing from evolutionary scientists Eörs Szathmáry and John Maynard Smith, he attributes the “human intelligence explosion” to our ability to form societies. “Individual humans aren’t very smart,” he said, “but when we come together, we can build rockets, perform organ transplants, and go to the Moon. That’s collective intelligence in action.”
In conclusion, Agüera y Arcas’ argument reframes the AI debate. Intelligence — whether biological or artificial — emerges from computation, cooperation, and evolution. Far from being an imitation of the human mind, AI might be a continuation of the same evolutionary process that produced intelligence itself. In this view, the line between artificial and natural intelligence is not just blurred — it may not exist at all.
