Do computers learn in the same way as people?

There are some important differences in how computers and people learn.  You can think of computer learning as similar to following a recipe. Computers learn by using rules – usually, the rules are written by humans, but in the case of AI, the computers can figure out the rules by being shown many examples. For example, AI algorithms can figure out how to recognize a cat because they have been shown thousands of pictures of cats, and thousands of pictures of things that are not cats, and they have been told which pictures include cats.  People do learn some things by rules (“Capitalize the first word of a sentence!”).  But people learn lots of things in other ways - through experience, including by making mistakes and figuring out what works and what doesn’t.

There are also a lot of things that are important for how people learn that aren’t important for how computers learn. Your computer doesn’t need sleep to learn, but you do! You also need to pay attention to information to remember it.  It also helps you to learn if you are curious about the information, if it’s important to you, or if it is associated with some type of emotion. These features don’t matter to a computer. A computer will learn what it is programmed to learn.

People are also better at transferring knowledge from one situation to another. If you’ve learned how to rollerblade, this can make it easier for you to ice skate.  Computers don’t transfer their knowledge in the same way.

There’s lots of interest right now in whether computers can learn in more human-like ways, and vice-versa.