AI learns language from distorted sources. This could change the way we humans speak – and think | Bruce Schneider

bBecause of the way they are trained, large language models only capture a portion of human language. They are trained in the written word, from textbooks to social media posts to our language as captured in films and television. These models have minimal access to the unscripted conversations we have face-to-face or voice-to-voice. This is the vast majority of language and an essential part of human culture.

There is a risk. The increasing use of large language models means that we humans are confronted with a lot more AI-generated text. We humans, in turn, will begin to adopt the linguistic patterns and behavior of these models. This will impact not only the way we communicate with each other, but also how we think about ourselves and what is happening around us. Our view of the world can be distorted in ways we barely begin to understand.

This will happen in many ways. One of the first impacts we could see is in simple expression, similar to how texting and social media have led to…