What is the difference between artificial intelligence and machine learning, and when would you use one over the other? originally appeared on Quora: the place to gain and share knowledge, empowering people to learn from others and better understand the world.
I’m more of a media and technology theorist than a technologist, so my answer to a question like this will be necessarily pedestrian.
Machine learning is really just the way a computer or algorithm establishes a way of doing something. So an autonomous vehicle will “watch” the decisions made by thousands of real drivers in order to “learn” how to drive. A bot will learn how to respond to a tweet by “reading” millions of responses to other tweets. Machines also learn through trial and error – iterating, repeating, and seeing what results they get. Machine learning is a little different than human learning, because machines are not looking for the reasons why things happen. They just care about the results. So if they learn that trading a stock ten seconds after 9:03am on alternate Wednesdays when it’s raining yields a profit, they will do that – without caring why the weather or time impacts anything.
Artificial intelligence is a phrase we use to describe the way computers think. There’s specific AI, like AI’s that can work in a very particular area (what we have now) and then there’s a theoretical concept called General AI, which would be computers that can solve problems across different sorts of activities. So we have AI’s that can imitate conversations with us, or create mashup art based on our instructions (cats flying to Mars) but not ones that go back and forth between all these kinds of things.