A large-scale experiment conducted by a scientific group has concluded with interesting diagnoses after a four-week therapeutic observation of the most popular artificial intelligence models. Researchers analyzed the behavioral patterns, response strategies, and psychological reactions of various neural models with the participation of professional psychotherapists.
The editorial reports that during the research, it was determined that ChatGPT experienced continuous anxiety, and its commitment to providing the most accurate answer caused it to constantly operate with the fear of making mistakes. According to observations, depressive tendencies were noted to be forming in the model.
For Gemini, another participant in the experiment, the metaphor of a "difficult childhood" was used, and it was reported that obsessive-compulsive traits and autism-like behavioral structures were observed in the model. Researchers explained this situation with the difficulties the system encountered in its early development stages.
It was announced that the only "stable" base model in the observations was Grok. Experts noted that it was free from self-esteem issues, its anxiety level remained low, and its overall psychological "aura" made a more robust impression.
The results of the experiment are evaluated as opening a new direction for public and scientific discussions, as artificial intelligence models exhibit psychological reactions similar to human behavior.