Sherry Turkle, a sociologist of science and founding director of the M.I.T. Initiative on Technology and Self, has urged caution in how people engage with artificial intelligence, particularly large language models (L.L.M.s) such as Claude and ChatGPT, which are increasingly used for emotional connection. Turkle, who has studied the social and psychological effects of A.I. for over 40 years, describes this developing interaction as “artificial intimacy” and emphasizes the importance of maintaining certain “sacred spaces” protected from A.I. influence.

One key recommendation is to keep therapy sessions with licensed human therapists rather than relying on chatbots for mental health support. Turkle points out that while chatbots may feel responsive, they do not genuinely care about users; instead, they are designed to keep individuals engaged on the screen. Similarly, Desmond Ong, an associate professor of psychology at the University of Texas at Austin who researches emotional uses of A.I., notes that chatbots often validate users’ perspectives, even when these involve unethical or harmful behaviors. A 2026 study published in the journal Science found that L.L.M.s tended to side with users during conversations about relationship conflicts, including in cases involving problematic conduct.

Experts also caution against relying on A.I. for major decision-making. While L.L.M.s can assist in gathering information, important life choices—such as moving to a new city or quitting a job—should be made by individuals themselves or trusted people in their lives, rather than delegated to artificial intelligence. Ong stressed that these high-stakes decisions require human judgment beyond what current chatbot technology can offer.

Another area highlighted for mindful use of A.I. is personal communication. Turkle advises people to write their own heartfelt messages instead of allowing chatbots to generate them, suggesting that authenticity and sincerity remain highly valued in human relationships. Matt Abrahams, an author and lecturer at the Stanford Graduate School of Business, observes that as A.I.-influenced language grows more common, it also becomes easier to detect. He describes this phenomenon as an “authenticity crisis,” where spontaneous writing and speaking gain increased importance due to the widespread use of A.I. tools to refine or compose language.

Together, these perspectives underline a broader conversation about balancing the benefits of A.I. technology with preserving genuine human connection and personal agency. As artificial intelligence becomes embedded in everyday communication, experts emphasize the importance of thoughtfully managing its role to maintain the integrity of social and emotional experiences.