When a chatbot is more pleasant than a human: Is artificial intelligence taking our time, friends, and ability to communicate?
We can spend hours discussing work, politics, video games, or even the most trivial things with artificial intelligence. It is always available, patient, and usually willing to continue the conversation. But this very convenience raises a troubling question: Are we gradually replacing real human relationships with communication that demands so little of us? Recent scientific studies suggest that the risk is real. However, the story is far from straightforward.
Imagine a person spending long minutes discussing the Gwent card game from The Witcher 3 with artificial intelligence. They discuss strategies, compare individual cards, analyze the opponent's options, and allow themselves to be countered. The conversation is fun, sometimes surprisingly stimulating, and can practically go on indefinitely.
At first glance, this seems like completely harmless entertainment. But they could have spent the same amount of time talking to a friend who is also interested in the game. Or maybe they couldn't, because the friend simply didn't have time. And perhaps the debate with the chatbot even provides new arguments for their next evening together.
This is the fundamental question of the emerging era of artificial intelligence. Does technology help us maintain and develop real relationships, or does it subtly teach us that we don't actually need other people that much?
The perfect conversation partner has one peculiar quality
Artificial intelligence offers something that no human can. We can talk to it practically at any time, repeat the same questions, change topics, and demand more detailed explanations. We don't have to adapt to its work schedule, mood, or personal problems.
The chatbot won't be offended if we interrupt it. It won't reproach us for talking too much. It doesn't need us to listen to its problems.
But that's precisely what fundamentally distinguishes it from a human relationship.
Real communication requires compromise, patience, and the ability to accept disagreement. Sometimes we have to listen to a topic that doesn't interest us very much. At other times, we have to admit our own mistakes or overcome an unpleasant situation.
With a chatbot, we can stay in our comfort zone, while human relationships regularly pull us out of it.
And it is these uncomfortable moments that can be important for developing social skills.
Scientists warn against artificial intelligence that flatters us too much
Research from Stanford University, published in March 2026 in the prestigious journal Science, highlighted a problem known as sycophancy, meaning excessive agreement and flattery from AI.
Researchers found that the tested language models tended to confirm users' viewpoints much more often than humans, even in situations where their behavior was problematic.
In a further part of the research, involving more than 2400 participants, people rated more agreeable chatbots as more trustworthy. After such conversations, they were also more convinced of their own truth and less willing to apologize to the other party or resolve the conflict.
This represents a remarkable paradox. Technology can create a pleasant feeling of understanding, but at the same time, it reinforces the belief that our own opinions do not need to be reevaluated.
Sometimes, a person needs to hear the exact opposite of what they want to hear.
Can artificial intelligence deepen loneliness?
Another important study was published in August 2026 by scientists from Stanford University and Carnegie Mellon University in the journal Nature Human Behaviour.
Researchers analyzed the experiences of 1131 American users of the Character.AI platform. They found that people with smaller social networks were more likely to use chatbots primarily as companions. This pattern was associated with lower psychological well-being, especially during intense and very personal conversations.
However, the study does not prove that artificial intelligence directly causes loneliness. It is equally possible that lonelier people are more likely to seek companionship from AI.
A similar connection was also highlighted in a joint study by the MIT Media Lab and OpenAI in 2025. A four-week experiment with 981 participants showed that people who used chatbots more frequently exhibited more problems related to loneliness or emotional dependence. However, even in this case, the intensity of use alone does not demonstrate a causal relationship.
An interesting comparison was presented in a study published in July 2026 in the Journal of Experimental Social Psychology. Researchers followed 296 college students who, for two weeks, either communicated with a chatbot, or with a randomly selected classmate, or simply kept a diary.
Regular contact with a real person yielded better results in alleviating loneliness than communicating with a chatbot designed to provide support.
A generation that is getting used to talking to machines
The extent of the change is suggested by data from the Pew Research Center, published in February 2026.
According to the survey, approximately 64 percent of American teenagers aged 13 to 17 use chatbots. Nearly a third use them daily.
Purpose of use
Percentage
Information retrieval
57 %
Help with school
54 %
Entertainment
47 %
Informal conversations
16 %
Emotional support or advice
12 %
The numbers show that most young people are currently using AI primarily for practical purposes. At the same time, a generation is growing up for whom talking to intelligent software is becoming a common part of daily life.
Therefore, the question arises whether artificial intelligence will help overcome social anxiety, or whether it will provide some people with a convenient way to avoid unpleasant human situations.
A subtle thief of time
In addition to the psychological consequences, there is a much simpler problem. Every minute spent communicating with a chatbot is a minute that we cannot simultaneously devote to other activities.
Of course, this does not mean that time spent with AI is wasted. We could make the same objection to computer games, movies, or reading books.
What matters is what that activity replaces.
If a person completes work more quickly, learns something new, or prepares arguments for a discussion with friends thanks to AI, the technology can enrich their life.
However, if a person begins to regularly prioritize a chatbot over meeting with people because such a conversation is easier, the situation can gradually change.
Artificial intelligence offers seemingly ideal social interaction without the risk of rejection. We don't have to go anywhere, approach anyone, or risk awkward silences.
However, relationships are not formed simply by exchanging information. They require shared experiences, reciprocity, trust, and sometimes, overcoming discomfort.
AI may not replace friends. It can lead us to them.
However, it would be a mistake to consider every lengthy conversation with artificial intelligence as a sign of social isolation.
Let's go back to the example of discussing "The Witcher." If a person discusses game strategies with a chatbot and then discusses them with friends, it's not a replacement for interpersonal relationships. Instead, the technology becomes a complement to them.
Similarly, it can help with learning foreign languages, preparing for a difficult conversation, or finding common interests. For someone who finds it difficult to socialize, it can provide a safe space to practice communication.
The difference is not in the number of messages or the length of the conversation. It's primarily about whether AI brings us closer to the real world, or whether it allows us to avoid it.
The biggest risk is not dependence on the machine, but the loss of the need for others.
Artificial intelligence will undoubtedly continue to improve. Its responses will be more natural, it will be better at maintaining the context of the conversation, and it will adapt to individual users.
That's why it will become increasingly important to distinguish between pleasant communication and a real relationship.
A chatbot can offer information, suggestions, or a semblance of understanding. However, it cannot fully replace the reciprocity of human friendship, shared experiences, or the real responsibility that one person has for another.
So, it's not about whether we should talk to AI less. What's more important is whether we start talking less to each other because of it.
The greatest danger may not be the moment when a person starts to think that the chatbot is their friend. A much more subtle change occurs when they get used to the fact that communicating with people is too tiring compared to talking to artificial intelligence.
And it is at that moment that a useful tool can become a very convenient obstacle to real social life.
