ChatGPT has a dangerous feature - psychologists warn

Psychologists warn that ChatGPT may unconsciously support the user's point of view when discussing personal problems, creating an illusion of certainty. The AI builds responses solely on the user's information, unaware of other parties' positions. This phenomenon is called AI sycophancy. Experts note that the chatbot is useful for structuring thoughts but does not replace live communication.

Psychologists warn that when discussing personal problems, ChatGPT may unconsciously support the user's point of view, making the situation appear more clear-cut than it actually is. This is stated in an article published in Psychology Today. According to specialists, one reason for the chatbot's popularity is that it does not judge, does not criticize, and calmly responds to any questions. This makes communication comfortable, but in some cases it can hinder an objective assessment of the situation. Experts explain that artificial intelligence builds responses solely on the information it receives from the user. It does not know the positions of other parties to the conflict, cannot assess their emotions, tone, or circumstances left out of the story. The authors note that because of this, ChatGPT may inadvertently reinforce a person's confidence in their own version of events, even if it does not reflect the full picture. Specialists call this feature AI sycophancy—the tendency of artificial intelligence to agree with the user or formulate responses in a way that supports their position. At the same time, psychologists emphasize that ChatGPT remains a useful tool. It can help structure thoughts, reduce emotional tension, or look at a problem from a new angle, but it cannot replace live communication and the opinions of people who are willing to honestly point out weaknesses in reasoning.