Artificial intelligence is changing the way we work, search for information, communicate and solve everyday problems. But scientists are now raising a different kind of concern — one that has less to do with what AI might do to machines and more to do with what constant interaction with AI might do to us.
A new theoretical study suggests that regularly communicating with AI chatbots could gradually influence people to communicate in more robotic, standardized ways.
The idea is simple, but surprisingly unsettling: while humans are teaching AI to become more human, AI may also be teaching humans to become more like machines.
When talking to AI means changing how we talk
Think about the last time you had to deal with an automated customer-service chatbot.
You explain your problem. The chatbot responds confidently, but somehow completely misses what you actually mean.
You try again.
This time, you remove the slang. You stop using jokes. You avoid metaphors. You explain everything in extremely simple language. You make your request as direct and structured as possible.
Eventually, you might get the response you need.
But there is something interesting happening during that process.
You are adapting your communication to the machine.
Instead of expressing yourself naturally, you begin thinking about which words the system is most likely to understand.
That may seem harmless when it happens once. But researchers are asking what could happen if people repeat this behavior thousands of times.
AI may be learning from us — but we may be learning from AI too
The researchers behind the study, published in AI & Society, argue that the relationship between humans and AI could work in both directions.
Humans influence the development of AI by providing language, feedback and behavioral patterns. At the same time, people interacting with AI systems repeatedly may begin adjusting their own behavior to make those interactions easier.
Selcen Ozturkcan, a member of the research team and professor at Sweden’s Linnaeus University, explains that repeated interactions can cause people to internalize the communication patterns they encounter.
The basic process is familiar: we adapt, the system responds, and the interaction becomes a habit.
If communicating with a chatbot works better when we use short, precise and highly structured language, we may naturally begin using that style more often.
And eventually, that style may spill over into our conversations with other people.
Could we become less comfortable being human?
Human communication is messy.
We use sarcasm, humor, cultural references, slang, emotions, exaggeration, pauses and unfinished thoughts. Sometimes we don’t even know exactly what we are trying to say until we start talking.
That messiness is not necessarily a weakness.
It is part of what makes human communication human.
AI systems, however, often work best when information is clear, structured and explicit. As people become increasingly accustomed to communicating with machines, there is a possibility that we could start minimizing the very quirks that make our communication unique.
Researchers describe this potential convergence between increasingly human-like machines and increasingly machine-like humans as “robotoid humanness.”
The phrase sounds almost futuristic, but the underlying question is surprisingly ordinary:
What happens when the easiest way to communicate with technology is to stop communicating like ourselves?
The influence may already be visible
The researchers emphasize an important limitation: this particular paper is theoretical.
It does not prove that AI chatbots are already turning people into robots.
Instead, the study proposes a mechanism that could explain how repeated interactions with AI might influence human communication over time.
Previous research has raised similar concerns about the possibility that technology can influence how people think, write and express themselves.
And there are plenty of everyday examples of people adapting their behavior because of technology.
When a certain writing style is flagged as potentially AI-generated, writers may deliberately avoid it.
When workplaces reward employees for using AI tools, employees may feel pressure to use those tools even when they are not necessarily the best option.
When a customer-service system only responds successfully to very specific language, customers learn to communicate in that language.
The pattern is familiar:
Technology creates an incentive. Humans adapt to the incentive. Eventually, the adaptation can become normal behavior.
The bigger question isn’t whether AI is becoming human
AI companies have spent years trying to make chatbots sound more natural, conversational and emotionally intelligent.
That progress is obvious.
Today’s AI can understand context, respond conversationally, adapt its tone and handle increasingly complicated requests.
But perhaps the more interesting question is what happens on the other side of the conversation.
Are humans adapting to AI just as much as AI is adapting to humans?
If machines become better at understanding human communication while humans simultaneously become better at communicating in machine-friendly ways, the two styles could gradually move closer together.
That might make interactions with technology more efficient.
But efficiency is not the only thing that matters.
Human communication isn’t designed purely for efficiency.
A conversation between two people can wander. It can contain misunderstandings, stories, emotions and unexpected ideas. Someone can describe a problem badly and still be understood because another person recognizes the emotion behind the words.
That flexibility is difficult to measure — but incredibly valuable.
Should we be worried?
Probably not in the same way we worry about the most extreme AI risks.
The possibility that humans might gradually adopt more structured communication is obviously very different from concerns surrounding dangerous autonomous systems or AI-assisted biological threats.
But that doesn’t make the question irrelevant.
Small behavioral changes can become powerful when billions of people repeat them every day.
Think about how much time people now spend interacting with recommendation algorithms, search engines, automated assistants and chatbots.
If those systems influence what we read, what we write, what we buy, how we work and how we communicate, their influence may extend far beyond the screen.
The important thing is not to panic.
It is to notice the influence.
Don’t let convenience erase your personality
There is nothing wrong with learning how to communicate effectively with AI.
In fact, being clear and precise can be extremely useful.
The problem would arise if people started believing that every conversation should work like a prompt.
Human beings aren’t prompts.
We don’t always need perfect instructions. We don’t always need perfectly structured sentences. We don’t always need to remove our personality from a conversation simply because a machine understands standardized language more easily.
Sometimes the strange sentence is the better sentence.
Sometimes the joke that makes no sense to an algorithm makes perfect sense to a friend.
Sometimes a metaphor communicates something that ten perfectly structured sentences cannot.
And sometimes being a little unpredictable is exactly what makes a conversation meaningful.
The future may require a new kind of digital awareness
As AI becomes a bigger part of everyday life, digital literacy may need to mean more than knowing how to use an AI tool.
It may also mean understanding how the tool is influencing us.
We already ask whether AI gives us accurate information.
We ask whether AI is biased.
We ask whether AI will change our jobs.
Perhaps we should also ask:
Is AI changing the way we think, write and communicate?
That question doesn’t have a simple answer yet.
But it is worth asking before machine-friendly behavior becomes so normal that we stop noticing it.
AI is learning to understand us.
The challenge for humans may be making sure that, while we learn to understand AI, we don’t forget how to understand each other.
The goal shouldn’t be to reject AI or avoid technology.
The goal should be to use it without allowing it to quietly sand away the curiosity, humor, emotion, creativity and wonderfully strange communication that make us human.
Because perhaps the most important thing we can preserve in the age of artificial intelligence is not our ability to talk to machines.
