Contribution short abstract:

A computer providing effective psychotherapy is a controversial prospect. Objections to it are often framed in terms of its damaging effect on the "therapeutic alliance," the interpersonal "active ingredient" of therapy. Drawing on original fieldwork in Australia, I consider this claim.

Contribution long abstract:

The prospect of a computer providing effective psychotherapy is both a tantalizing technological feat and a potentially disturbing rupture in human-human relations. It is also an idea almost as old as modern computing itself: Joseph Weizenbaum’s famous ELIZA software in the 1960s– a crude but surprisingly engaging Rogerian– finds echoes in contemporary automated CBT programs and AI chat-bots. These most recent efforts to automate the talking cure have been applauded, but also met with skepticism and alarm about the encroachment of machines on what is meant to be a delicate and subtle interpersonal process.

In this paper I draw on original ethnographic fieldwork conducted with psychologists in Australia, where such “e-mental health” projects receive funding from the federal government as public health initiatives. Reflecting professional orthodoxy, many informants object to such automated interventions on the grounds that the “therapeutic alliance” is absent or perverted– the nebulous “active ingredient” of psychotherapy characterized by mutual recognition and interpersonal rapport. Still others, often those researching and championing these automated interventions, claim that the alliance’s primacy has been overstated or question it altogether.

I argue that disagreements around this contested professional idiom highlight deep and unresolved questions about psychology. Specifically, about therapy as the exclusive domain of the skilled listener, the necessary physical limits of the session, and the unique ability of humans to heal one another.