The notion that humans are flawed machines waiting to be upgraded with AI is a misguided comparison, says Sarah O’Connor, a prominent AI researcher.
A flawed analogy
The human brain is often described as a machine, a notion that has been around for centuries. Marvin Minsky, a pioneer in artificial intelligence, famously referred to humans as “meat-based computers” in his 1967 book, “Computation: Finite and Infinite Machines”. However, this analogy is more limiting than liberating, according to O’Connor. She argues that this comparison invites us to see ourselves as sub-optimal alternatives to AI agents, a notion that ignores the complexity and uniqueness of human cognition.
While AI systems are designed to excel at specific tasks, our brains are capable of processing a vast array of information, from abstract concepts to sensory experiences. We can reason, create, and empathize in ways that AI systems currently can’t replicate.
The limits of AI
Despite the rapid progress in AI research, there are still significant limits to what AI systems can achieve. They lack human intuition, creativity, and emotional intelligence, which are essential for making decisions in complex, real-world scenarios. “AI is good at answering questions, but it’s not good at asking them,” says O’Connor. “It’s not good at understanding the nuances of human behavior or the subtleties of language.”
What this means
So, what does this mean for our understanding of AI and its role in society? It means that we shouldn’t be looking to AI to “augment” or “improve” human cognition, but rather to complement it. By recognizing the unique strengths and limitations of both humans and AI, we can design systems that work together in harmony, rather than trying to replace or enhance one another. As O’Connor puts it, “We’re not machines, and we’re not just machines with a bit of soul added on. We’re something entirely different, and that’s what makes us human.”
