The AI voice: polished, predictable, and problematic

Ask ChatGPT to help you write copy, and you might find yourself “navigating challenges”, “leveraging insights”, or “delving into” topics. These aren’t just stylistic quirks, they’re part of a growing linguistic shift driven by artificial intelligence. And for many, it’s starting to feel a bit... off.

As AI tools become embedded in our daily communication, whether through writing assistants, chatbots, or predictive text, they are shaping the way we speak. But the language they generate is often polished, Americanised, and very corporate.

The AI-generated language

AI-generated language tends to favour certain words and phrases. “Utilise” often replaces “use”. “Folks” appears more frequently than “people”. And “impactful” is everywhere. These choices reflect the data these models are trained on, largely American, corporate, and formal in tone.

For British users, this can feel like a subtle erosion of linguistic identity. Spellings like “favour” and “realise” are often corrected to their American counterparts. Idioms and regional expressions are ironed out in favour of what the model deems “clear” or “professional.”

 Built-in bias

The dominance of US English in AI outputs isn’t accidental. Most large language models are trained on datasets that skew heavily toward American media, websites, and academic writing. As a result, British English is often treated as a variant… or worse, an error.

This has led to awkward moments in professional settings, where AI-generated content is flagged for being “too American” or “not quite right”. 

The backlash begins

As AI-generated language becomes more common, so does the backlash. On social media, users have begun mocking the “AI voice”, a tone that’s overly polished, emotionally distant, and riddled with buzzwords. Terms like “meticulous”, “delve”, “realm”, and “adept” are used up to 51% more frequently than in the three years prior researchers found.

In academia, students are facing scrutiny for using AI to draft essays, with some UK universities reporting a rise in disciplinary cases. While some praise AI for boosting confidence and productivity, others say it’s stifling creativity and homogenising expression.

A 2023 survey by The Verge found that while many people are curious about AI, a significant number are wary of its influence, particularly in creative and communicative fields.

Whose language wins?

The rise of AI-assisted communication raises deeper questions about linguistic power. Who gets to decide what “good” language looks like? And what happens when a machine’s version of clarity becomes the global standard?

If we continue to use AI to shape how we communicate, we may need to push back, not just to preserve regional dialects and cultural nuance, but to protect the diversity of human expression itself.

There’s no denying that AI is here to stay. It’s already shaping our culture, including the way we speak and write. But the choice is ours: we can either lean fully into the confident, perfectly sophisticated voice AI offers, or we can strike a balance, one that leaves room for the natural, unedited human.