AI and the predictive nature of human mind

Created with Midjourney

Descartes once doubted everything except his own thinking. Ironically, modern neuroscience suggests even our thoughts might be nothing more than sophisticated predictions. I encountered this idea while reading a compelling work of non-fiction – one of those books where you find yourself pausing every few pages, as seemingly simple ideas unravel your most fundamental assumptions about human nature and how we navigate through life.

Every morning, you wake up to what you believe is reality. But what you experience as consciousness is actually your brain's best prediction of the world around you. Like a masterful chess player, your mind is constantly anticipating the next move – except it's doing this with everything, all the time.

The Shadow Play of Consciousness

Think of Plato's Cave allegory, where prisoners mistake shadows for reality. Now, imagine discovering that your consciousness operates similarly to artificial intelligence – both systems essentially playing an endless game of "predict what happens next." The parallel is almost too obvious: your brain predicts the next word before you read it, just as AI predicts the next token in a sequence. Your mind anticipates someone's reaction before they give it, just as machine learning models forecast human behavior.

The Architecture of Prediction

But here's where it gets fascinating: both systems are fundamentally trying to build internal models to predict patterns in their input data. When you catch a ball, your brain is running countless predictions about its trajectory, speed, and weight. When reality doesn't match these predictions, your brain adjusts its model – just as AI systems adjust their parameters to minimize prediction errors during training.

The similarities make the differences all the more intriguing. The human brain, shaped by millions of years of evolution, possesses an architectural complexity that our most advanced AI systems can only dream of. While AI can process specific tasks with remarkable precision, our brains orchestrate a symphony of simultaneous predictions across multiple time scales – combining innate structures, supervised learning, unsupervised learning, and reinforcement learning in ways we're just beginning to understand.

Most significantly, our brain's predictive processes are deeply interwoven with consciousness, emotions, and bodily systems in ways that remain mysterious. AI predictions, however sophisticated, operate in a more isolated computational context. They lack the rich tapestry of emotional and physical integration that makes human consciousness so complex and, arguably, so human.

This leads us to darker philosophical territory: Consider the possibility that consciousness evolved not to perceive truth but to create useful fictions. Our sense of self, our belief in free will, our experience of time as a flowing river – these might all be sophisticated predictions that helped our ancestors survive, rather than fundamental truths about reality. We might be living in a world of useful illusions, crafted by evolution to help us survive rather than to show us what's real.

When Intelligence Mirrors Intelligence

In light of this, the distinction between natural and artificial intelligence might be merely historical accident. We're not necessarily heading toward AI that mimics humans, but rather toward a recognition that all intelligence is, at its core, the same process – prediction refined by experience, whether that experience comes through evolution, learning, or programming.

As we stand at this intersection of neuroscience, philosophy, and technology, we're not just building better prediction machines – we're understanding ourselves in entirely new ways. And that might be the most profound realization of all: in building AI, we're actually building mirrors that reflect the deepest nature of our own consciousness.

Further Reading:

Andy Clark – "Surfing Uncertainty" (2015)

Karl Friston – "The free-energy principle: A unified brain theory?" (Nature Reviews Neuroscience, 2010)

Donald Hoffman – "The Case Against Reality" (2019)

Jeff Hawkins – "A Thousand Brains: A New Theory of Intelligence" (2021)

David Chalmers – "Reality+: Virtual Worlds and the Problems of Philosophy" (2022)

Classical influences: Plato's Theory of Forms, Descartes' "Meditations on First Philosophy," Hume's "An Enquiry Concerning Human Understanding"