• Reading Level 5

Technology Journalist of the Year Aged 14-18: Winner

Beeline Reader (learn more) uses subtle colour gradients to help you read more quickly and efficiently.

“Outputting good title…” Is AI already changing how we express ourselves?

Tom, James Gillespie’s High School, UK

It’s no secret that the use of ‘generative artificial intelligence’ – AI that can generate new content, such as text – has become commonplace in schools since the advent of ChatGPT in late 2022. Since then, the technology has only grown in use elsewhere too, from producing product adverts on Amazon, to holding difficult conversations on dating apps like Tinder. But while generative AI, and the deserving media attention it’s garnered, are relatively new, other forms of AI have actually been around for quite some time – and they’re changing how we write.

Not by composing entire essays for us, mind you. This form of AI is a lot more basic, but it’s built on the same principles as the newer models: it’s trained on huge amounts of data, and decides what makes good writing accordingly. It can analyse vast bodies of text and suggest improvements, grammar corrections, and can grade the text based on the atmosphere you’re trying to create, or the register you’re trying to speak in. This ability is achieved by algorithmically comparing the new text with its training data.

Of course, this refers to programmes like Grammarly and Microsoft Editor, so-called ‘writing assistants’ – or so they were called. Grammarly, formerly marketing itself as such, changed its description to ‘AI-powered’ within a month of ChatGPT launching. Perhaps that’s a recent example of AI influencing the way we express ourselves.

In any case, it certainly isn’t the first. Microsoft Editor comes pre-packaged into Office 365’s version of Word Online, a service used by three hundred and forty-five million paying users. As of 2014, five hundred billion Office documents were created yearly, and four trillion emails were sent in that same timespan – and the watchful eye of Microsoft Editor sees, annotates, and edits all of them. In comparison, Grammarly’s thirty million-strong user base seems tiny, but the number is nothing to sneeze at. The takeaway: AI could touch everything we read.

But… what does it really do? In the case of Microsoft Editor, it approximates the tone of your writing into one of three broad categories – formal, professional, casual – and makes judgements to deliver that tone best. Scrutinising this very article, the AI provided suggestions not just as to how the document could be edited, but why; it was providing justifications, referring to maximising impact and conciseness. Technically speaking, it was trying to fit the text into its algorithms. That same process is used by Grammarly.

It’s clear that AI has been prevalent since long before the generative boom of the 2020s, and it has a big capacity to affect our self-expression within the constraints of algorithms. While it remains to be seen what effect this has on shaping our language, it’s worth considering that some of the current era’s most culturally significant authors – the likes of J. K. Rowling and Stephen King – have used Microsoft Word when writing. Online journalists and contributors to large, widely-read publications like the New York Times often do too. Now, the latest generation is being raised on it – your school probably uses Word! So next time you’re reading something, think about the extent to which the language is the author’s own, or the output of an algorithm.

Print Friendly, PDF & Email