Researchers from the University of Michigan indicate that communication with artificial intelligence can contribute to a deeper understanding of personality traits.
A recent study has shown that interacting with generative AI models, such as ChatGPT, Cloude, and LlaMa, can help us understand human personality with remarkable depth and accuracy. These models were able to analyze written speech and assess personality traits with a degree of accuracy that surpassed traditional methods, and in many cases, even exceeded the accuracy of the opinions of close friends.
Researchers at the University of Michigan based their study on an analysis of the diaries and personal histories of 160 participants. The study revealed that AI assessments closely aligned with the participants' self-perceptions. This accuracy extended beyond personality traits to include emotional state, stress levels, social interaction patterns, and even indicators of potential psychological problems.
Professor Aidan Wright of the University of Michigan explained that the findings confirm that personality is naturally reflected in everyday language and personal narratives, even without conscious self-expression. Despite the encouraging results, the researchers emphasized the need for further studies to verify the accuracy of these assessments across different demographic groups and to examine the influence of factors such as age, gender, and ethnicity.
The study showed that artificial intelligence analyses can predict realistic aspects of individuals’ lives, including their psychological state, their response to stress, their social behavior, and even the presence of diagnosed psychological disorders or previous requests for help.
For his part, Professor Chandra Shripada, a professor of philosophy and psychiatry at the University of Michigan, pointed out that these results support a well-established idea that language holds deep keys to understanding individual differences in personality and temperament, noting that artificial intelligence has made analyzing personal texts faster, more accurate, and easier.
In contrast, the study authors acknowledged a number of limitations, most notably the research’s reliance on participants’ self-reports, the lack of direct comparison of AI results with assessments from friends or family members, and the unclear impact of different demographic factors.
Scientists also pointed out that it is not yet known whether artificial intelligence and humans rely on the same signals when assessing personality, or whether these models will be able in the future to outperform self-assessments in predicting important life paths, such as relationships, education, health, and career success.
In the study’s conclusion, psychologist Colin Wise of the University of Pittsburgh confirmed that artificial intelligence has proven its ability to reliably detect personality traits through everyday language, opening new horizons for a deeper understanding of the human psyche
