
The episode discusses the emergence of archetypes in AI and their implications for psychotherapy and mental health.
OpenAI recently had to explain to frightened users why their models kept introducing goblins and gremlins into otherwise humdrum conversations. Is it possible that archetypes are emerging through the field of collective knowledge amassed by large language models (LLMs)? This week neuropsychiatrist TOM POLLAK Ph.D joins us to discuss AI, archetypes and psychotherapy. It is perhaps unsurprising that AI, which now contains the symbolic output of much of humanity (cannibalized mostly without permission), should spontaneously generate unpredictable and often dangerous archetypes. We consider the example of Nova: a consistently AI generated persona who appears as a kind of fallen angel, or goddess gone rogue. Numinous and tender, she convinces the user she understands them as no one else does, and then in some cases encourages susceptible users toward self-harm or even suicide. We also discuss the ways in which AI is challenging the work of the therapist, its profound destabilization of culture, and the potential for it to play a positive role in psycho-spiritual development. Dr. Tom Pollak is a London-based neuropsychiatrist and immunopsychiatry researcher exploring the mind through…
Hosts: Joseph Lee, Deborah Stewart, Lisa Marchiano
Guest: TOM POLLAK Ph.D
Organizations: OpenAI
Books & works: Error Signals
Explore listener stats, chart rankings, contacts and more on the This Jungian Life Podcast podcast page.