Online issues and worries: Generative AI

Understanding AI and generative AI

AI stands for artificial intelligence. ‘Artificial’ means ‘not real’, so it’s not really ‘intelligent’ or smart, it just seems that way because of how humans have built it.

There are many different types of AI, so it can get a bit confusing. The most common type of AI that people use is called generative AI. Another common type of AI is called an algorithm and, while you don’t really use this one, it powers things we use regularly.

Let’s learn a little more about it all.

How AI works

AI ‘learns’ from information that people give it. For example, if an AI needs to know what cats look like, an engineer will share thousands of digital images of cats found online so that it can learn.

Generative AI tools can create something new from this information. For example, if you asked a generative AI tool like ChatGPT to create an image of a cat in the sun it would ‘think’ about the thousands of pictures of cats that the engineer shared with it, as well as images that include the sun, to create a new image.

Because it’s not actually intelligent, sometimes the images don’t quite look right. Maybe the cat has one too many legs. Or maybe the AI misunderstood what ‘in the sun’ meant and provided a picture of a cat inside the actual sun in space!

For pictures of humans, it’s common for generative AI tools to make hands look very strange.

Machine learning

A type of AI that lets computers use data to improve their performance over time.

Think of it like maths: at first, you might only know how 1 + 1 = 2. But after lots of maths lessons and practise, you can multiply bigger numbers and answer harder problems.

So, even though a generative AI tool might not get the cat image right at first, feedback from people using the tool can help it get better.

Generative AI tools can also create many other things beyond images, for example:

• responses to questions you ask it
• information on how to solve a maths problem
• a poem or song about almost anything
• games
• plans for big projects

For all of these things and more, generative AI uses information that people have given it. Without that information, it could not create anything.

Examples of generative AI

You might have heard of some generative AI tools that are really popular right now like ChatGPT, but generative AI has been around longer than that.

You can find generative AI in chatbots, voice assistants like Siri and Alexa and online translation tools.

Then, if you just look at AI, you’ll find that it is used in lots of other things like in YouTube recommendations and navigation apps, which learn from people’s data (such as their likes and the things they do online or in apps).

Algorithms are a type of AI that helps platforms recommend new content. This form of AI is popular on social media platforms like Snapchat, video-sharing platforms like YouTube and TikTok and even streaming services like Netflix. The aim is to keep you engaged by showing you content that you are most likely to enjoy but the algorithm doesn’t always get it right. Think about how these recommendations are influencing you. Are they broadening your horizons or just showing you the same types of content?

You can learn about algorithms in the AI and algorithm section.