When people talk about artificial intelligence, it's tempting to think that it means computers can think (a little) like humans. Perhaps unsurprisingly it turns out that's not quite the case.
This video from New Scientist explains how machines are designed to be capable of solving problems that humans can -- just not quite in the same way. Instead, it all boils down to data, probability and a lot of code written by humans -- but that doesn't stop them being incredibly, undeniably useful. [YouTube]