In a recent clinical encounter with one of the authors (AP), a young boy with autism insisted that his mother hated him—because ChatGPT said so. After asking whether a parent who sets limits must dislike their child, he interpreted the bot’s confident, literal response as truth. By the time the boy arrived at the clinic, this exchange had already reshaped his affect, relationships, and risk. Encounters like this are increasingly common, and many adolescents now present with beliefs—and sometimes safety concerns—influenced by generative artificial intelligence (AI).
Radically different