AI chatbots have been designed to function as confidantes, lovers, advisors, friends, playmates and collaborators of one sort or another. Whilst imaginative diversity of uses frequently exceeds the intentions of developers the business model of commercial chatbots is premised on building trust in the reliability of the technology and the relationships it affords. With their on-demand access and relatively low subscription costs, chatbots are also responding to a large reservoir of unmet mental health need, ranging from everyday companionship to psychotherapy – most commonly based on CBT or Rogerian counselling techniques. Psychodynamic understanding of how humans and chatbots interact is still at an early stage, but it is increasingly clear that we need new conceptual frameworks to characterise this distinctive mode of human-machine relating.

Drawing on our work developing AI companions with both communities of lived experience and psychoanalytic psychotherapists, we will introduce you to a series of chatbots designed to offer everyday companionship, and for use in a care eco-system. We ask what capacities you would look for in your ideal AI companion, what limitations you could accommodate and what use might you make of it? What are the opportunities and challenges that chatbots pose for care professionals and users? How far can they be trusted? what are the implications for designers and developers?