Imagine being able to understand the world around you, adapt to new situations, and make decisions based on your surroundings. That’s what animal intelligence is all about. But how close is ChatGPT, a highly advanced language model, to achieving this level of intelligence?
To explore this question, let’s start with a reference point: the cat. Cats have a cerebral cortex, the outer layer of their brain, which processes sensory information, controls voluntary movement, and supports learning and decision-making. It’s estimated that a cat’s cerebral cortex contains hundreds of millions of neurons and trillions of connections between them. Interestingly, ChatGPT’s latest models have roughly the same number of parameters, or artificial ‘neural connections,’ as a cat’s cerebral cortex.
However, that’s where the similarities end. A cat’s cortex is made of living neural tissue, which reshapes and adapts with every experience. ChatGPT’s artificial neural network, on the other hand, stays fixed once it’s trained. It can adapt for a moment, but it doesn’t truly learn from experience.
So, how do we measure intelligence? Is it just about the number of neurons or connections? Or is it about how those neurons interact with the world around them? A cat’s intelligence is embodied, meaning it’s connected to its physical environment and can feel, see, and touch the world. ChatGPT’s intelligence, by contrast, is linguistic, existing solely in the world of words and meanings.
But here’s the fascinating thing: intelligence has meaning only within the world it’s part of. A cat excels in a world of survival, where its senses and adaptability are crucial. ChatGPT, on the other hand, shines in a world of language and knowledge. And while it may not be able to smell, taste, or touch, it can process and analyze vast amounts of information with ease.
So, how close are we to achieving something we could call superintelligence? Would it require a brain the size of a human’s? The answer, much like the question, is complex. But one thing is certain: the line between human and artificial intelligence is blurring, and it’s up to us to define what intelligence truly means.
As we continue to push the boundaries of AI, we may uncover new ways of measuring intelligence. We may find that ChatGPT and other AI models are not just mimicking human intelligence but creating their own unique forms of cognition. And who knows? Maybe one day, we’ll be able to create AI that can truly learn from experience, adapt to new situations, and make decisions based on its surroundings.
For now, the journey continues, and the question remains: how close is ChatGPT to animal intelligence?