Definition
You recognise your friend's face in a second. How does your phone do it?
Natural intelligence is the intelligence of living beings, which grows in a brain through experience. Artificial intelligence is intelligence shown by machines, which humans design and which learn from data or follow knowledge given to them.
In simple words
A child learns faces by seeing family members every day, with eyes and a brain. A phone learns faces from thousands of photos, with a camera and a computer chip. Both end up recognising faces, but they learn in very different ways, and the phone can do only that one job while the child can do thousands.
Why it matters
Comparing the two helps you see what AI is good at and where it falls short. It also explains why AI needs so much data: it does not have years of life experience, a body or common sense to fall back on.
Key terms
- Natural intelligence
- The intelligence of living beings, which develops in a brain through growth and experience.
- Artificial intelligence
- Intelligence shown by machines that people design and build.
- Neuron
- A nerve cell in the brain that passes signals to other nerve cells. The human brain has about 86 billion of them.
- Artificial neural network
- A computer model made of simple connected units, loosely inspired by neurons in the brain.
- Training data
- The examples a machine learns from.
- Common sense
- Everyday knowledge about how the world works that people pick up without being taught, such as "water makes things wet".
Explained in detail
What "artificial" really means
"Artificial" here means made by people, not fake. Artificial light from a bulb is real light; it was simply made by a human invention and not by the sun. In the same way, artificial intelligence is real problem-solving ability that has been built into a machine by people. The word tells you where the intelligence came from, not whether it is genuine.
The brain and the computer
The human brain contains about 86 billion neurons, each connected to thousands of others. It runs on about 20 watts of power, roughly what a small light bulb uses, and it learns all the time, from the day we are born. A large AI system runs on thousands of powerful chips in a data centre and uses vastly more electricity to learn. Artificial neural networks borrow one idea from the brain: many simple units connected together, where the strength of each connection changes during learning. Beyond that, they work very differently from real neurons.
How each one learns
A child learns from a few examples, with the help of curiosity, play, people who explain, and a body that can touch, taste and move. Once a child has seen a few cats, they recognise a cat in a photo, a cartoon or a shadow. Most AI systems need thousands or millions of examples, and they learn only from the kind of data they are given. A face-recognition model never tastes food, never falls off a bicycle and never asks "why?". This is a big reason why natural intelligence is so flexible and artificial intelligence is usually so narrow.
Where machines are ahead, and where people are
Machines are far ahead in speed, memory and tirelessness. A model can compare a face with millions of saved faces in a second and do it all night. People are far ahead in common sense, understanding feelings, moving through a messy world, and transferring what they learned in one area to a completely different one. A person who learns to ride a bicycle finds a scooter easier. An AI trained to play chess cannot play cricket at all.
| Question | Natural intelligence | Artificial intelligence |
|---|---|---|
| Where does it come from? | Growth, experience and the brain | Design and data |
| Examples needed to learn | Often a few | Often thousands or millions |
| Has feelings? | Yes | No |
| Uses one skill in a new area? | Easily | Rarely |
How it works
| Natural intelligence | Artificial intelligence | |
|---|---|---|
| Where it lives | In a brain, made of living cells | In computers, made of chips |
| How it learns | From experience, senses and other people | From data, examples and rules people give it |
| How many tasks | Thousands, and it switches easily | Usually one kind of task |
| Speed and scale | Slow at huge amounts of data | Can check millions of items quickly |
| Feelings and common sense | Yes | No proven feelings or common sense |
Try it yourself
A child and a machine learn the same thing
Both learn to recognise a cat. Watch how differently they get there.
Natural or artificial?
Decide whether each example shows natural or artificial intelligence.
A dog finds a hidden ball by smell
A maps app predicts traffic on your route
A child learns a new word from a story
A phone unlocks when it sees your face
A crow bends a wire to pull out food
Common mistakes
- Thinking 'artificial' means fake. The abilities are real; they are built by people.
- Thinking AI works like a brain. Some AI is inspired by brains, but it works very differently.
- Thinking AI is smarter than people in everything. It is faster at narrow tasks, not wiser in general.
Check yourself
Summary
- Natural intelligence grows in living brains through experience.
- Artificial intelligence is built by people and learns from data or rules.
- AI is fast and tireless at one kind of task; people are flexible across thousands.
Finished reading?