Are we slowly writing like ChatGPT, or is ChatGPT progressively writing like us? When I was at school, I wrote love letters for other boys courting girls. Some relationships, and perhaps even marriages, may have been founded on words from my pen!😜 That is a story for another day. I became a wordsmith of sorts, largely because I was an obsessive reader (still am). I read anything I could get my hands on, including novels, newspapers from front to back, religious books, autobiographies, magazines, and even adult content. If it had words, I read it. Over time, without consciously imitating anyone, my vocabulary, writing style and command of language came to reflect what I consumed. Later in my professional life, colleagues sought me out for reports, strategies, manuals, policies, standards and all manner of official correspondence. Decades later, this experience has led me to an intriguing question. If reading what other humans wrote shaped how I write, could repeatedly reading, editing and refining ChatGPT's output eventually make us write more like AI, even when we are writing entirely on our own? Many of us increasingly use ChatGPT, Grammarly, QuillBot and other AI applications to draft, edit, paraphrase, polish and even “humanise” our writing. We read the outputs, revise them, read them again, and repeat the process, sometimes several times a day. Could repeated exposure to AI-generated language be quietly reshaping how we write, even when we are not using AI? Not by copying and pasting, but through the same subconscious absorption of vocabulary, syntax, rhythm and structure that shaped my writing decades ago. I suspect we may be witnessing the early stages of what I described in June as AI-mediated linguistic convergence: a gradual alignment of human writing with patterns commonly found in AI-generated text. And here is another fascinating possibility. As AI influences human writing, and some of that writing enters the datasets used to train future AI models, could we create a feedback loop in which humans and machines increasingly shape each other's language? When I first raised this question in my June article, it was a personal hypothesis. Since then, research published in Nature Human Behaviour has reported measurable reductions in linguistic diversity associated with AI-assisted writing. That does not yet prove the subconscious influence I am suggesting, but it makes the question even more interesting. I do not necessarily see this as a bad development. It may improve clarity, accessibility and the quality of written communication. But it could also influence individuality, creativity and the distinctive voices we develop through our experiences. I became a writer largely through reading what others wrote. Soon, we could learn more and more from writing produced by both humans and machines. Have you noticed yourself adopting expressions, vocabulary or sentence structures from ChatGPT, even when writing entirely on your own?