I’ve been thinking a lot and reading a lot and listening to a lot and talking a lot about artificial intelligence in general, and large language models in particular. I am a psychoanalyst and I run a small group practice. In the community of psychoanalysts and psychotherapists, many of us are incredibly suspicious or fearful of artificial intelligence. My colleagues fear that ChatGPT will take our jobs and will lead to severe mental health harms.
I’m struck as I talk to fellow clinicians by the incredibly shallow knowledge and incredibly strong opinions that many of us hold. My own perspective is unusual among psychoanalysts and may be born of the fact that prior to becoming a psychoanalyst I had a career in private equity investing. It never occurred to me to see artificial intelligence generally or ChatGPT specifically as a threat. Rather they seem like exciting tools that, like any tool, can be used for good or evil.
Many of the most overblown fears and concerns about artificial intelligence in the wider world have been imported without thought by my fellow clinicians. Concerns about hallucination, about agency, about alignment. I should say that many of my fellow clinicians have these concerns but don’t really understand what those concerns are when held by those deeply familiar with and knowledgeable about artificial intelligence.
Unquestionably, if you ask bad questions of ChatGPT, you will get bad answers. Also: if you hammer your thumb it will hurt. At the same time there is no doubt in my mind but that artificial intelligence, ChatGPT, and other large language models represent a tremendous opportunity.
In the coming days I hope to write a fair amount about these subjects to explore both the sources of our anxiety and the promises and perils of artificial intelligence for psychotherapy in general and psychoanalysis in particular. I will describe some of the various ways that I have attempted to use ChatGPT and HeyPi, and other artificial intelligence tools.
I will describe some of the ridiculous failures I’ve experienced and some of the really exciting and promising successes. I hope to share my enthusiasm and my humility on the subject.
I have only one strongly held opinion and that is that artificial intelligence will be used well by some and poorly by others. Beyond that, I really do feel quite humble.
