One of the contested differences between human language and a large language model is that humans have mental intentions. These intentions may lead to an external gesture in speech, but not always. However, this area of distinction has been blurred by LLMs that now have a deep reasoning mode. In this mode, the model will offer something similar to mental intentions before they give their answer.

However, there seems still to be one area that distinguishes human thought and intention and the output of an AI LLM. This is in the phenomenon known as Tip of the Tongue phenomenon. As its name suggests, human can often struggle to find a word that they need; they have some sense of what it is and they feel they are on the verge of remembering it- however, at that moment they can’t. This difficulty demonstrates that humans can have the concept in their minds but they don’t have the words to express it. In other words, humans can think outside language, whereas all the indications seem to suggest that LLMs cannot. Moreover, this ability gives us the freedom to think outside the box of language and come up with an originality that LLMs cannot.