These circulate heavily on TikTok, Twitter, and group chats:
“It depends.” Used jokingly or seriously, this is now a default nuanced answer.
“Great question!” People start messages with this ironically or earnestly
“Delve” has become an academic neon sign in the middle of every conversation flashing ChatGPT was here
A country-music AI generated song reached the top of the US Charts for the first time this week. “Walk my Walk”, by Breaking Rust ( an artist with no identity).
Introduction
Have you ever read a message and thought, “Wow, this sounds like it was written by a polite and well-read Roomba vacuum”? The sentences are spotless; the phrasing is oddly robotic. Ironically, it was written by a human, just one who’s apparently been binge-watching ChatGPT output and picked up the accent. For those who live and breathe words like mediators, arbitrators, neutrals, etc. it’s both fascinating and mildly terrifying since there is a dependence on nuance, tone, and the human touch that can’t be automated (yet). But what happens when your client sounds like a chatbot? Large language models (LLMs) have learned to write like humans by absorbing billions of our words / tokens, and now, it seems, we’re returning the favor. After ChatGPT went public, researchers noticed people sprinkling their speech with AI’s favorite buzzwords. It’s like the AI mimicked us, and we’re now mimicking the mimic. A feedback loop of linguistic masquerading.
Technology and Communication Changes
The idea that technology can change the way we communicate is not new, not by a long shot. History is full of examples of tools that reshaped human interaction and behavior:
So no, it’s not shocking that new technology is influencing our communication. But LLM-based AI is a different beast. Unlike the printing press or electricity, these AI systems engage with us on a deeply personal level, they literally speak our language, back to us and through us. When you’re drafting an email and your email client suggests a light-hearted “Got it, thanks!” or your phone offers “On my way!” before you even think to type it, that’s AI nudging your words. And guess what? Those little suggestions have started to slip into our everyday talk. Words and phrases that come from predictive text and chatbots , simple stock lines like “sounds good” or “let me check and get back to you” , are increasingly showing up in normal, human-to-human conversations. . I’ll admit it: I’ve absolutely caught myself saying, “Sounds good, I’ll get back to you,” in a conference call, basically quoting my email autoreply like it’s from the “Terminator” (I’ll be back). It’s convenient, sure. But it does make you wonder: if all your texts end with “👍🏻 no problem,” do you slowly become a more agreeable person, or just someone who’s lost the will to disagree?
At some point, the whole situation starts to feel like a feedback loop: humans create AI, AI generates highly probable language, humans adopt that language, which then further trains the AI, and so on. It blurs the line between who is the teacher and who is the student in our communication. As one scientist observed, it’s possible for AI to sway the vocabulary of entire populations on a large scale, and such shifts in language can end up reshaping how we think, our cultural norms, even public discourse itself. In other words, shift how people talk and you may shift how people think. These are the new “terms and conditions” of human language in the AI era: our words are no longer just ours; they’re part of an ongoing exchange with our machines.
Culture Codes
If this sounds a bit like sci-fi or social engineering, it kind of is, except it’s happening right now, in plain sight. We nod along with machine-guided communication norms without even realizing it. For professionals in mediation and arbitration, this might mean seeing parties interact in slightly more AI-influenced ways: perhaps both sides in a dispute come armed with meticulously worded statements that feel as if ChatGPT polished them. Maybe everyone’s being extra polite and reasonable (a relief, frankly!), but you might sense a certain homogeneity in tone that wasn’t there a decade ago. This is a new cultural script being co-written by humans and machines.
Model Behavior
What we’re really talking about is conditioning, training, if you will, happening on a societal scale. Social scientists even have a term for this kind of adaptive feedback: operant conditioning. The concept is simple: when a behavior or phrase seems rewarded (maybe it gets you a positive response, or just feels socially “correct”), you do it more. If using a certain eloquent word makes you sound smart and everyone nods with approval, you’re likely to use that word again. AI is adding another layer to this dynamic. It provides instant feedback and suggestions that shape our behavior in real time, almost like a high-tech Skinner box for language and social interaction.
Not sure? Well, consider some warnings from experts. One fear is that we might lose diversity in language and expression if everyone is converging toward the same AI-suggested style and vocabulary. “Humans could lose language diversity,” cautions Hiromu Yakura of the Max Planck Institute, describing a potential negative feedback loop where AI and humans just keep reinforcing the same patterns. In a worst-case scenario, we end up in an echo chamber of phrasing, a sort of monoculture of language. Yakura even used the term “core collapse” of language to imagine a future where nuance and variation collapse under a homogenized human-machine dialect. For professionals who thrive on nuance (hello, mediators!), that idea is downright scary.
Another concern is how AI might spread bias or misinformation and subtly narrow our thinking. Today’s social media algorithms already show the power of selective amplification: they boost certain content and bury other content to keep us engaged and scrolling. Tomorrow’s AI systems could easily generate floods of persuasive messages tailored to push emotional buttons or political agendas. It’s not far-fetched to think that authoritarian regimes or even savvy negotiators could deploy AI-generated language to sway public opinion, or an opponent’s opinion, in their favor. When every word we see or hear could be fine-tuned by AI for impact, maintaining an open and unbiased perspective becomes a real challenge.
Conclusion
Even though this sounds dystopian, there can be benefits to this strange new interplay between human and machine communication. It’s easy to imagine some positive social trends emerging. For example, AI-generated text often defaults to a very polite, measured tone (sometimes almost to the point of blandness, but still). Where human communication was historically terrible (think Twitter feuds, mediation of high conflict divorces, or nasty email threads), an AI’s influence might nudge us to be more courteous and careful with our words. Perhaps we’ll all get better at writing clearly and concisely too, after all, an AI can suggest how to tighten up a sentence or avoid an ambiguous phrase, essentially coaching us in real time. Some have speculated that AI could even improve our decision-making processes by offering logical, less emotionally charged framing of problems (like Nextlevelmediation.com). In conflict resolution, imagine an AI tool that helps parties phrase their demands in less confrontational language. If that technology “trains” people to communicate more constructively, that’s a win for everyone. The likely outcome is a co-evolution: we and our AIs will learn to live in a shared communicative space, each influencing the other.
For mediators, arbitrators, and other neutrals, all this is both a challenge and an opportunity. On the upside, if AI-driven habits make people’s everyday communication more polite or thoughtful, the job might get a little easier. I mean, who wouldn’t prefer a mediation opening statement that starts with “Thank you for your perspective, I appreciate it” instead of an aggressive tirade? If disputing parties, start mirroring an AI’s dispassionate tone, maybe discussions stay more rational and less inflammatory. On the downside, we must be vigilant that all this niceness isn’t just a facade. There’s a difference between authentic empathy and a neatly worded “I understand your position” that was suggested by a GPT-5 script. Mediators and arbitrators have developed keen ears for sincerity. They might soon need an equally keen ear for AI fingerprints in someone’s speech or writing. Is that apology at the bargaining table coming from the heart, or from a template?
In the end, human communication has always evolved with our tools. From quill pens to telephones to texting emojis, we’ve continuously redefined how we express ourselves. AI is just the latest tool that’s pushing our evolution forward. The key will ensure that through all these shifts we maintain the human element. A machine might help us talk more nicely, more efficiently, even more fairly but it’s our job to keep the conversation real.
Endnotes:
Author's Acknowledgement [1] “We [mediators] are victims of unintended consequences - a ‘revenge effect’ – a result unanticipated that is almost the exact opposite of the one we designed.” -...
By Peter AdlerWe are all shouting, but no one is listening. That can be said of American politics in general, but what can we do to restore constructive deliberation of important issues....
By Paul WahrhaftigBefore social media, the personals column found at the back of New York Magazine was where readers sought potential mates or more, often beginning their ads, as they were called...
By Arthur L. Pressman
