About this video
This short video explains that AI is teaching computers to do things that usually need human intelligence. Using a robot dog that learns to fetch, it shows how basic AI learns from lots of data, then explains generative AI, which learns from the internet to create new things but can make mistakes, so we must use it critically and responsibly.
You'll be able to
- explain what artificial intelligence is in my own words.
- describe how AI learns from lots of examples, like a robot dog learning to recognise toys.
- explain what generative AI does and name problems it can have.
Key words
- Artificial intelligence (AI)
- Teaching computers to do things that usually need human intelligence, like recognising objects or understanding speech.
- Data
- The information, like lots of pictures, that AI learns from.
- Generative AI
- A kind of AI that learns from the internet and creates new things, like answers, stories or pictures.
Transcript
You've probably heard a lot about AI recently, but what is it, really? AI, or artificial intelligence, is when we teach computers how to do things that usually require human intelligence, like identifying an object, understanding human speech, and even talking.
But how do you teach a computer to "learn" and "think"? Well, it's kind of like when you train a pet to do tricks. Think of AI as a robot dog that you're training to fetch a toy. At first, your robot dog might not even know what a "dog toy" is! So that's the first thing you need to teach it.
You show it lots of pictures of "dog toys" so that it learns to recognize them quickly and easily. It might make some mistakes at first, but with each correct answer, it gets a reward. Over time, the dog's recognition improves, the same way you improve at any task the more you do it.
And once the robot dog gets really good at recognizing what a dog toy is, you can move on to teaching it the next step of playing fetch – running after the toy and bringing it back to you. This is basic AI, which learns by analyzing lots of data – in this case, pictures of dog toys and how to play fetch.
But there's something even more complex, called generative AI, which is more like a creative robot that can improvise when it comes to making art or writing. Unlike the robot dog, the creative robot learns from everything on the internet – videos, text, photos, you name it!
When you ask it a question or give a hint, it uses what it has learned from the internet to create new things – like answers, stories, or even pictures. But here's the catch: that creative robot can sometimes make mistakes. And it doesn't know the difference between good and bad.
It doesn't know if what it creates is helpful or hurtful. It can't always tell facts from fiction, or know where its information comes from. It might even use someone's work without crediting them. It just sucks up all the random or not-so-random information that's floating out there on the internet.
That's why we, as humans, have to think hard and be critical about the information we receive from AI. We have to know what it does well and what it doesn't do well. As we use generative AI in our lives, remember that we have to be smart and responsible with how we use it.
Video © Khan Academy, used under CC BY-NC-SA. Watch on Khan Academy. Puzzicle is not affiliated with Khan Academy.