How AI Is Changing The Words We Use Daily

While many worry about AI sounding more human, a surprising new study suggests the opposite is also happening: we are starting to sound more like AI. Research indicates that the formal, slightly academic language common to chatbots like ChatGPT is seeping into our everyday conversations. Words like "delve," "meticulous," and "underscore" are becoming more frequent, subtly shaping how we communicate.

The Study: AI's Lexicon in Human Speech

A recent study from Florida State University has uncovered a measurable shift in our language patterns since the launch of ChatGPT in 2022. Researchers analyzed over 22 million words from unscripted science and technology podcasts, which are typically built around informal and natural discussions. The findings were clear: words that are stylistic hallmarks of large language models (LLMs) are now appearing far more often in human speech than they did before the AI boom.

Interestingly, the study noted that people are not just adding these words to their vocabulary; they are actively choosing them over simpler synonyms. For instance, a speaker is now more likely to say "delve" instead of "explore," or "underscore" rather than "accentuate." It seems we are unconsciously absorbing the specific stylistic tics of AI-generated text, much like picking up a catchphrase from a popular show. This trend isn't isolated to podcasts, as other research has found the same thing happening among academics.

A Linguistic Feedback Loop

This might seem like a harmless evolution of language, but researchers warn of a potential feedback loop with worrying consequences. As we adopt AI's vocabulary and produce content using that language, future AI models will be trained on this human text that is already AI-influenced. This could accelerate linguistic change in a way that homogenizes our speech, making everyone sound like the same bland digital assistant.

Of course, technology has always influenced language. Texting gave us abbreviations and emojis, while social media popularized hashtags and new slang. However, this shift feels different. Instead of adding new, creative elements to our vocabulary, it involves replacing more freewheeling and evocative words with a more formal, standardized lexicon.

While we are not on the verge of a language apocalypse, the way we speak directly affects how we think. If the language we use to frame ideas and make decisions is subtly shaped by the patterns of an AI, it could influence our thought processes in ways we don't yet fully understand or appreciate. Further research will need to delve into this phenomenon to underscore its long-term impact.