The day of reckoning has arrived, and it’s not quite how The Terminator envisioned it. Instead of taking over the world, chatbots are taking over our vocab. And no, your usage of internet lingo isn’t just a symptom of how chronically online you are. It’s an indication of a larger socio-lingo evolution, fueled by the digital age.Â
Max Planck Institute for Human Development released a report, showcasing how the usage of LLMS like ChatGPT and Gemini are actively altering how people speak and communicate with one another–particularly, in terms of our word choice.Â
After the release of ChatGPT, a group of German researchers analyzed thousands of academic talks on YouTube and conversation podcasts, finding a noticeable spike in the usage of AI’s most frequently used words. They discovered that the word “delve”, for instance, one of the many overused terms by AI, became more commonly used colloquially.Â
Now I’m not trying to be a fearmongerer. While studies like this signal a possible lexical shift, suggesting AI’s potential to influence our collective vocabulary, reality tells us otherwise. It’s an impossible task to quantify the true imprint chatbots are leaving on human communication, especially as we’re amidst the early years of its metamorphosis. There’s no doubt it will take its shape, regardless of calls for intervention, if that provides any solace.Â
According to American linguist and author Adam Aleksic, we internalize the words subconsciously, carrying them over into our writing and speaking in the workplace, school, and other settings. This is especially problematic because LLM models are consuming data that are themselves the outputs of such models, which causes the machines to reinforce their own outputs and perpetuate fabricated information and misinformation.
With “transformer architecture,” also known as deep learning architecture, which the AI models from the past few years are based off of, these models are operating off numerical representations that are coded into tokens, and producing language by predicting the next token in the sequence. It’s safe to say, it’s quite literally taking the humanity out of language.Â
If the trend continues, there may be homogenization of language in our future, and inevitably, the loss of language, which raises concerns about language and cultural preservation.Â
While it’s unrealistic to resist the inevitable waves on human behavior AI will have in the future, we must make an effort to preserve our individual, personalized ways of speaking, however informal, peculiar, or uneloquent it may be!