Is ChatGPT going to change the world?

As I write about technology for a living, I’m used to seeing small leaps in technological progress. This year’s hot new phone has a slightly better camera than last year’s, for instance. Or a new computer processor has been invented that can crunch numbers slightly faster than what came before. 

But occasionally, there are things that will shock me. At the end of last year, a company called OpenAI released a truly jaw-dropping demonstration of a new AI technology that feels almost like the technology has skipped a few generations in between. It feels equivalent to going straight from the Wright Brothers to a Boeing 747. And I genuinely think it is going to change the world. 

The technology is called ChatGPT—though you can be forgiven for mistaking it for witchcraft, magic or divine intervention. So what exactly is it? 

ChatGPT is a computer program that can understand and respond to human language. It is trained on a lot of text, so it can understand what people are saying and respond in a way that sounds like a real person. It can be used for things like having a conversation or finishing a sentence. Even this sentence that you are reading now, was written by ChatGPT. 

That’s right—to write the paragraph above, I asked ChatGPT to “Write a paragraph containing a basic explanation of ChatGPT in simple terms, that at the end reveals that the paragraph itself was written by ChatGPT,” and it came up with that explanation of itself, all on its own. Because the system was built by basically feeding it what feels like a significant chunk of the internet, it seemingly knows how to talk about everything—from politics and sport to science fiction. 

What’s amazing is that it seems so eerily human in its ability to understand questions and answer them competently. “Write a story about Boris Johnson in the style of Roald Dahl,” you can ask it, and it’ll spin up a pretty good children’s story. “Write a list of reasons why London is better than New York,” and it’ll explain how London is more historic and has better public transport. “Write an episode of Star Trek where the Chuckle Brothers join the Starship Enterprise,” and it’ll write whatever bizarre crossover episodes you can dream up. It can even write songs and limericks too. 

One area where I’ve found it enormously useful is that it can write computer code. This particularly amazed me as unlike the English language, when you write code it has to be very precise and carefully structured—but when I asked it to write me a program that would put some data into a database for me, in a particularly complex way, within seconds it generated something that would have taken me hours to do manually. 

It isn’t, however, completely human-like. For example, it sometimes generates things that sound true, but are just nonsense. And it can’t do maths—I asked it what is the square root of 717,409 and it told me the answer is “approximately 838.8” (the answer is actually 847).

In both cases, this is because of the way the AI learns through pattern recognition. There’s no real intelligence there—it is just regurgitating something based on what it has "read" before. So it is better at some tasks than others—it can write a brilliant, sincere-sounding condolence message (umm, not that I’ve used it in this way…yet), but its ability to rhyme leaves a lot to be desired. 

I think the best way to think of ChatGPT is (ironically) like a calculator. It doesn’t replace the human completely—but it does make doing the sums much quicker. In the same way, it’s easy to imagine in the not too distant future every time we write something, we may have an incredibly smart AI assistant helping us get started by generating a draft, on whatever topic we’re writing about. In fact, Microsoft is reportedly already planning to build the technology into its Office apps. 

And this is why I’m pretty convinced that ChatGPT—or something like it—is going to change the world. It’s already hugely impressive, and what exists at the moment is just a demonstration. Once the same AI is built into other apps, and once it can connect to the actual, live internet to learn more, it will become even more sophisticated. Our computers will no longer be just our word processors, but our writing partner too. 

So now I’m just hoping that my editor doesn’t decide it could write a pretty decent Reader’s Digest technology column without me. 

Read more: 5 Questions about the future of tech, answered

Read more: Top tips for spotting deepfakes online

Keep up with the top stories from Reader's Digest by subscribing to our weekly newsletter

*This post contains affiliate links, so we may earn a small commission when you make a purchase through links on our site at no additional cost to you.