Abstract As artificial intelligence (AI) becomes integrated into mental health care, foundational questions about psychotherapy gain renewed urgency. What makes treatment “real”? Which therapeutic elements are essential for genuine change? This article addresses these questions through a conceptual examination of two leading treatments for depression: psychodynamic therapy (PDT) and cognitive-behavioral therapy (CBT). While comparative studies of these treatments often report equivalent outcomes, echoing the “Dodo bird verdict” that “everybody has won, and all must have prizes”, this article traces their core differences. The goal is not to promote dogmatic allegiance to “pure” forms of therapy, but to identify distinctive ingredients for each tradition, explore conceptual and methodological “troubles” that obscure them, and discuss their potential adaptation to AI. A side-by-side analysis of canonical texts highlights three features distinctive to PDT: (1) attention to past experiences; (2) open-ended exploration of fantasy life; and (3) focus on the therapeutic relationship itself, through transference, countertransference, and immediate psychic truth. These features are difficult to capture in short-term psychodynamic treatments, which dominate comparative research. They are also particularly challenging for digital and AI adaptation. Practical directions for research are outlined (e.g., studying contrasting strategies such as distraction, alongside clinical outcomes beyond symptom relief, such as mentalization). The article concludes with an examination of the possibilities and limitations of AI-based psychotherapy. While many CBT components appear amenable to automation, the human connection emphasized in PDT (and by many CBT practitioners) remains difficult to replicate. The article explores whether AI could, nevertheless, support such relational depth.
Yaakov Ophir (Sat,) studied this question.