Join any Zoom call, walk into any lecture hall, or watch any YouTube video, and listen carefully. Past the content and inside the linguistic patterns, you\u2019ll find the creeping uniformity of AI voice. Words like \u201cprowess\u201d and \u201ctapestry,\u201d which are favored by ChatGPT, are creeping into our vocabulary, while words like \u201cbolster,\u201d \u201cunearth,\u201d and \u201cnuance,\u201d words less favored by ChatGPT, have declined in use. Researchers are already documenting shifts in the way we speak and communicate as a result of ChatGPT \u2014 and they see this linguistic influence accelerating into something much larger.<\/p>\n
In the 18 months after ChatGPT was released, speakers used words like \u201cmeticulous,\u201d \u201cdelve,\u201d \u201crealm,\u201d and \u201cadept\u201d up to 51 percent more frequently than in the three years prior, according to researchers at the Max Planck Institute for Human Development, who analyzed close to 280,000 YouTube videos<\/a> from academic channels. The researchers ruled out other possible change points before ChatGPT\u2019s release and confirmed these words align with those the model favors, as established in an earlier study comparing 10,000 human- and AI-edited texts<\/a>. The speakers don\u2019t realize their language is changing. That\u2019s exactly the point. <\/p>\n One word, in particular, stood out to researchers as a kind of linguistic watermark. \u201cDelve\u201d has become an academic shibboleth, a neon sign in the middle of every conversation flashing ChatGPT was here. \u201cWe internalize this virtual vocabulary into daily communication,\u201d says Hiromu Yakura, the study\u2019s lead author and a postdoctoral researcher at the Max Planck Institute of Human Development.<\/p>\n \u201c\u2018Delve\u2019 is only the tip of the iceberg.\u201d<\/p>\n But it\u2019s not just that we\u2019re adopting AI language \u2014 it\u2019s about how we\u2019re starting to sound. Even though current studies mostly focus on vocabulary, researchers suspect that AI influence is starting to show up in tone, too \u2014 in the form of longer, more structured speech and muted emotional expression. As Levin Brinkmann, a research scientist at the Max Planck Institute of Human Development and a coauthor of the study, puts it, \u201c\u2018Delve\u2019 is only the tip of the iceberg.\u201d<\/p>\n AI shows up most obviously in functions like smart replies, autocorrect, and spellcheck. Research out of Cornell<\/a> looks at our use of smart replies in chats, finding that use of smart replies increases overall cooperation and feelings of closeness between participants, since users end up selecting more positive emotional language. But if people believed their partner was using AI in the interaction, they rated their partner as less collaborative and more demanding. Crucially, it wasn\u2019t actual AI usage that turned them off \u2014 it was the suspicion of it. We form perceptions based on language cues, and it\u2019s really the language properties that drive those impressions, says Malte Jung, Associate Professor of Information Science at Cornell University and a co-author of the study.<\/p>\n This paradox \u2014 AI improving communication while fostering suspicion \u2014 points to a deeper loss of trust, according to Mor Naaman, professor of Information Science at Cornell Tech. He has identified three levels of human signals that we\u2019ve lost in adopting AI into our communication. The first level is that of basic humanity signals, cues that speak to our authenticity as a human being like moments of vulnerability or personal rituals, which say to others, \u201cThis is me, I\u2019m human.\u201d The second level consists of attention and effort signals that prove \u201cI cared enough to write this myself.\u201d And the third level is ability signals which show our sense of humor, our competence, and our real selves to others. It\u2019s the difference between texting someone, \u201cI\u2019m sorry you\u2019re upset\u201d versus \u201cHey sorry I freaked at dinner, I probably shouldn\u2019t have skipped therapy this week.\u201d One sounds flat; the other sounds human. <\/p>\n For Naaman, figuring out how to bring back and elevate these signals is the path forward in AI-mediated communication, because AI is not only changing language \u2014 but what we think. \u201cEven on dating sites, what does it mean to be funny on your profile or in chat anymore where we know that AI can be funny for you?\u201d Naaman asks. The loss of agency starting in our speech and moving into our thinking, in particular, is what he is worried about. \u201cInstead of articulating our own thoughts, we articulate whatever AI helps us to articulate\u2026we become more persuaded.\u201d Without these signals, Naaman warns, we\u2019ll only trust face-to-face communication \u2014 not even video calls. <\/p>\n We lose the verbal stumbles, regional idioms, and off-kilter phrases that signal vulnerability, authenticity, and personhood<\/p>\n The trust problem compounds when you consider that AI is quietly establishing who gets to sound \u201clegitimate\u201d in the first place. University of California, Berkeley research<\/a> found that AI responses often contained stereotypes or inaccurate approximations when prompted to use dialects other than Standard American English. Examples of this include ChatGPT repeating the prompt back to the non-Standard-American-English user due to lack of comprehension and exaggerating the input dialect significantly. One Singaporean English respondent commented<\/a>, \u201cthe super exaggerated Singlish in one of the responses was slightly cringeworthy.\u201d The study revealed that AI doesn\u2019t just prefer Standard American English, it actively flattens other dialects in ways that can demean their speakers. <\/p>\n This system perpetuates inaccuracies not only about communities but also about what \u201ccorrect\u201d English is. So the stakes aren\u2019t just about preserving linguistic diversity \u2014 they\u2019re about protecting the imperfections that actually build trust. When everyone around us starts to sound \u201ccorrect,\u201d we lose the verbal stumbles, regional idioms, and off-kilter phrases that signal vulnerability, authenticity, and personhood.<\/p>\n We\u2019re approaching a splitting point, where AI\u2019s impacts on how we speak and write move between the poles of standardization, like templating professional emails or formal presentations, and authentic expression in personal and emotional spaces. Between those poles, there are three core tensions at play. Early backlash signals, like academics avoiding \u201cdelve\u201d and people actively trying not to sound like AI, suggests we may self-regulate against homogenization. AI systems themselves will likely become more expressive and personalized over time, potentially reducing the current AI voice problem. And the deepest risk of all, as Naaman pointed to, is not linguistic uniformity but losing conscious control over our own thinking and expression. <\/p>\n The future isn\u2019t predetermined between homogenization and hyperpersonalization: it depends on whether we\u2019ll be conscious participants in that change. We\u2019re seeing early signs that people will push back when AI influence becomes too obvious, while technology may evolve to better mirror human diversity rather than flatten it. This isn\u2019t a question about whether AI will continue shaping how we speak \u2014 because it will \u2014 but whether we\u2019ll actively choose to preserve space for the verbal quirks and emotional messiness that make communication recognizably, irreplaceably human.<\/p>\n