How Chatbots Are Changing Human Language and Conversation

Summary

Chatbots are beginning to leave recognizable marks on human language, according to three Northeastern University experts. They point to the growing use of words such as “delve,” “essential” and “insight,” along with metaphors involving tapestries and symphonies, as possible stylistic fingerprints of AI. Such vocabulary may be temporary, while deeper grammatical changes usually take generations. Terra Blevins argues that large language models trained heavily on English may also encourage English-like subject-verb-object ordering in languages with more flexible word order, potentially reducing nuances that word arrangement can convey. AI influences conversation in both directions. In research discussed by Blevins, ChatGPT mirrored users’ wording and function-word choices more strongly than users mirrored the model, although people still adapted to the chatbot at rates broadly similar to human conversational alignment. Experts also describe a backlash: some writers avoid em dashes and AI-associated vocabulary because those features have become perceived “bot tells.” Beyond style, Sofia Teixeira says chatbots can help people organize thoughts and rehearse difficult conversations, potentially supporting agency. However, their constant agreeableness and instant regeneration may deprive people of the disagreement and repair that human relationships require. Overreliance could weaken self-regulation and conflict-handling, while simulated empathy may create an “echo chamber of one.”