Comparing human intelligence to artificial intelligence is meaningless, as the fundamental nature of intelligence in these artificial systems is fundamentally different from our own. Human intelligence cannot be reduced to mathematical equations or data processing as is the case with artificial intelligence. We communicate, understand and learn in complex ways, governed by emotions and social relationships, and these aspects cannot be programmed…
A question answered by Neil Lawrence, Professor of Machine Learning and Artificial Intelligence at the University of Cambridge.
What makes humans stand out from other living creatures or even inanimate matter? Many would say it’s intelligence. But the rise of intelligent machines is putting that belief to the test. The companies behind these technologies, like ChatGPT and its competitors, talk about achieving “artificial general intelligence,” the attainment of a level of intelligence comparable to humans in a variety of tasks.
Does this rapid advance in AI make our human intelligence less special? Neil Lawrence, a professor of machine learning at the University of Cambridge, doesn’t think so. In fact, he argues that we should abandon the concept of “artificial general intelligence” altogether. In his new book, Atomic Humans: Understanding Ourselves in the Age of Artificial Intelligence , Lawrence argues that a deeper understanding of our own intelligence and how it is fundamentally different from artificial intelligence can help us make the most of both.
The Atomic Human: Understanding Ourselves in the Age of AI is Professor Neil Lawrence’s latest book, which explores the complex relationship between human intelligence and artificial intelligence in our time. In this book , Lawrence argues that direct comparisons between human and artificial intelligence can be misleading and even harmful, and offers a different perspective on how to understand and harness both types of intelligence in a complementary way.
In this interview, Lawrence explains why he thinks both human and artificial intelligence are misunderstood, why there is no point in comparing them, and why we ultimately need a deeper understanding of intelligence.
What do you think about the tendency to compare artificial intelligence to human intelligence?
Lawrence believes that most of these discussions are meaningless and irrelevant. Artificial intelligence is fundamentally different from our own intelligence. It is absurd to talk about this kind of intelligence as if it were similar to our own. But as humans, we tend to humanize things because that is what helps us understand and communicate.
There was a tradition at one time of newspapers publishing formulas for beauty, made up of a bunch of ridiculous numbers, to try to explain what beauty was. Talking about intelligence in the same way, as if it were orderable and measurable, is also ridiculous. We don’t talk about “artificial general beauty,” so why talk about “artificial general intelligence”?
Are we any closer to a satisfactory definition of intelligence?
In his book, Lawrence uses the term “intelligence” a lot, but he tries to shift toward “information processing” as an alternative idea. We can now look at precise definitions of intelligence based on information, because information is the foundation. Large distributed systems process information, and so does the culture around us. We are a distributed intelligence.
Do you have an example of this type of distributed intelligence?
Social insects are an example of distributed intelligence. Lawrence describes observing ants in southern Italy, how they build multiple nests, distribute their tasks in an astonishing way, and make decisions about where to go to gather resources by communicating with each other in complex ways. Any reasonable definition of intelligence must include information-processing elements.
What does this tell us about the nature of intelligence in general?
Ants are information-processing systems, but so is the entire ecosystem around us. Information is stored in DNA, and plants, animals, bacteria, and archaea respond to changes in the environment and make decisions that affect multiple generations. The AI we’ve created may be closer in the way it processes information to ants than to humans.
What makes us see human intelligence as something unique?
Lawrence points out that the nature of our intelligence is also distributed, but in a way that makes communication between us limited, while our computational power is vast. This gives each of us a sense of self, but makes it difficult for us to grasp the nature of distributed intelligence or systems that operate over very long timescales, such as millions of years.
Can AI systems like ChatGPT ever reach human-level intelligence?
Lawrence is skeptical about claims of artificial general intelligence. He thinks it’s dangerous to talk in these terms because we don’t have a good definition of intelligence. But despite his skepticism, current AI represents an information revolution, changing the way we communicate with machines.
What is the current AI lacking?
One important element is motor skills. Everything we’ve achieved so far with computers has centered around mental tasks, like playing chess or solving equations. But the abilities that people enjoy, like playing games or providing emotional support, are still difficult for machines to replicate, especially motor skills.
What makes human intelligence unique and impossible for machines to emulate?
Lawrence argues that our unique characteristics, such as awareness of our limitations, death, loss, and betrayal, are what make human intelligence different and beautiful. These experiences and meanings embedded in our lives are what make us and our intelligence special.
