In her 2024 song Slim Pickins, Sabrina Carpenter sighs over her lover’s literary shortcomings: “This boy doesn’t even know the difference between there, their, and they are.” Her dry delivery says it all – she’s tired of dating guys who don’t know basic grammar.

It looks like she was on to something. According to a new study from MIT’s Media Lab, we should all be worried about losing our linguistic edge – thanks to AI.

Researchers split participants into three groups and asked each to write SAT-style essays. One group used ChatGPT. Another used Google. The third relied on nothing but their own brains. The ChatGPT users “consistently underperformed at neural, linguistic and behavioural levels.” Meanwhile, the “brain-only” group demonstrated the highest neural connectivity. They were also more engaged, more curious and felt more satisfied with what they’d written.

The dream AI companies like to sell us is that tools like ChatGPT make us faster, smarter and more productive. But this study suggests the opposite: leaning too hard on AI might actually be blunting our brains. 

So, as more individuals and brands turn to ChatGPT to compose, correct and smooth over their words, are human-made errors –  the typos and the clunky phrasing; all the things we used to dread – becoming more desirable? Maybe they’re not embarrassing anymore. Maybe they’re just … human. 

AI is making people question their ability to perform even basic tasks. A panelist on a recent talk said, “I feel like AI is fostering a culture of insecurity,” they said. “I do know how to formulate sentences, I’ve been at school for 15 years, I can write coherently, but I’m starting to lack trust in my own ability.” 

In an age hyper-fixated on productivity and optimisation, products like ChatGPT neatly meet the growing demand for fast and convenient outputs. In exchange, we begin to surrender our inherent ability to think for ourselves, turning to the artificial to express our human thoughts – raw feelings, sanitised but grammatically correct.