On AI, Human Intelligence, and Computer Outputs: The Airplane and the Bird
Researchers have used this analogy to argue that AI (the airplane) doesn't need to work like nature (the bird) to be useful. When we're looking at functional outputs (flight) as our definition of success, that may be acceptable. But what about when you're asking AI to understand your internal world? The difference starts to matter significantly, as the internal mechanics, much like a bird and a plane, are completely different; and the outputs can be considerably different when linked to the deeper processes inside.
Can an airplane get from location A to B quickly and accurately? Yes. But what if you want to know what flight feels like, or whether you should become a pilot? What if you wanted to feel your first tumble, wobble, and then first steps and then run and then the first time you hit the air, and the freedom and excitement that comes with that? What if you want to imagine diving through the depths, swooping through wind and clouds, taking in the sights below, pinpointing a target, and diving with precision? What if you wanted to understand ancient migratory patterns, instincts around where to fly, or darting to avoid predators?
You wouldn't want to ask an airplane to explain that. You'd want to ask a bird.
"Artificial Intelligence" May Be a Misleading Title
"Artificial intelligence" as a label challenges us to consider the definition of intelligence. A chatbot is technically a Large Language Model (LLM). LLMs are trained on enormous amounts of text, nothing more to learn from. (Text which was controlled by the people who had access to tools to produce it. In other words, a very curated edition of text; more on that in a later post.) LLMs learned statistical patterns in language. They can look at text, analyze it for patterns, and provide outputs accordingly.
You've removed everything in the human brain but language, nothing but words. LLMs did not grow up being held by a caregiver, and did not experience a litany of stresses, frustrations, bumps, embarrassments, joys, and relationships that changed them. LLMs do not have desires, feelings, sensory experiences, or personal memories. They do not experience being cold, embarrassed, proud, or in love. They only know these experiences through our words.
What Is Human Intelligence?
What we call intelligence arises from an enormously complex system with multiple brain modules working in parallel and interacting simultaneously. Consciousness is the felt experience of that, and is considered greater than the sum of its parts. Neuroscientist Antonio Damasio spent decades arguing that you cannot separate thinking from feeling, that the body and its emotional signals are not noise in the system: they are the system. Consciousness, he proposed, isn't a feature the brain added on top. It's what it is all doing, simultaneously and synergistically.
An overview of what the brain does, beyond just language:
A developmental process traced through trial and error with a brilliant mechanism to "wire together what fires together" (with all its emotional ups and downs) and synaptically prune away what is not used
A body interacting with the world, sending constant feedback to the system
Vision, hearing, touch, smell, taste, proprioception and interoception
Emotions… so many emotions
Motivations and drive towards growth
Social interconnectedness of various forms
Memory formed through lived experience and stored as patterns
Conscious awareness, defined as the felt experience of something
Brain systems that evolved for survival over millions of years, shaped by forces we don't even fully understand
We do not, and have never, had a fully accurate theory to adequately explain how those pieces combine to create intelligence or consciousness. As Stephen Hawking always pointed out, we probably couldn't understand it with the cognitive machinery we've been given.
The Output Problem
If you define intelligence as the ability to achieve a goal via problem-solving and pattern recognition, then perhaps you could argue an LLM is intelligent. If you define intelligence as an embodied experience which encompasses all the interwoven systems mentioned above: a felt experience with conscious, analytical, emotional and social elements, and a sprinkle of something mystical we can only wonder about, then you would definitely see that an LLM is lacking in major ways. It is the ultimate thin slice of what a brain is capable of.
We humans are biased and sadly easily impressed by visible outputs rather than the process beneath. Think of your biggest jerk colleague who can spout off great-sounding advice but lacks the wisdom, discernment, self-awareness, or mature judgement underneath. Much like that jerk, language is not the whole mind; it is simply the mind's most visible output.
LLMs can be impressively capable yet fundamentally different from a human mind.
A Category Difference, Not a Competition
A bird and an airplane both fly. Same output, completely different thing happening underneath. AI produces language the way a plane produces flight: through engineering that mimics the surface result without replicating the underlying nature.
We are not slower, messier versions of a language model. We are something else entirely.
Keep that in mind when asking for advice.
Subscribe to the Village Psychology newsletter for weekly psychology tools that help children feel safe, seen, and supported. GET IT HERE