Humans are able to bond emotionally with AI - sometimes more strongly than with humans

People are capable of experiencing a sense of emotional closeness to artificial intelligence - and in some cases even stronger than when communicating with other people. This is the conclusion reached by psychologists from the University of Freiburg and the University of Heidelberg. The results of the study are published in the journal Communications Psychology.

The two online experiments involved 492 people. The participants engaged in text conversations in which they discussed personal and emotionally significant topics such as important life events, friendships and interpersonal relationships. They were given answers either by another person or by an AI language model. Some of the participants knew who they were talking to and some did not.

The study showed that if the participants did not know that they were talking to an AI, the level of perceived closeness was comparable to a human. Moreover, in emotionally intense conversations, the AI evoked even stronger feelings of closeness than real conversations.

The key factor turned out to be the degree of self-disclosure. AI responses contained more personal information and emotional openness, which increased feelings of trust and engagement. Humans, as the authors note, tend to behave more cautiously when meeting an unfamiliar interlocutor.

However, the effect changed dramatically if the participants were told in advance that their interlocutor was an artificial intelligence. In this case, the sense of intimacy decreased, and the participants responded with less elaboration and put less effort into the conversation.

The authors note that the results open up possibilities for the use of AI in the fields of psychological support, care, education and counselling - especially as a "low threshold" tool for people with limited social contacts. At largely the same time, the study also points to risks.

According to the researchers, people can form emotional bonds with AI without realising it, making clear ethical and regulatory frameworks necessary. Without transparency, such systems could be used to manipulate users' emotional engagement.

The scientists emphasise that the future of AI as a "social actor" depends on how it is designed and regulated - whether it becomes a useful complement to human relationships or a tool to replace them.