Leonard Bickman's article on the future of artificial intelligence (AI) in psychotherapy research paints an encouraging picture of the progress to be made in this field. We support his perspective, but we also offer some cautionary notes about the boost AI can provide. We suggest that AI is not likely to transform psychotherapy research or practice to the degree seen in pharmacology and medicine because the factors that contribute to treatment response in these realms differ so markedly from one another, and in ways that do not favor advances in psychotherapy. Despite this limitation, it seems likely that AI will have a beneficial impact, improving empirical analysis through data-driven model development, tools for addressing the limitations of traditional regression methods, and novel means of personalizing treatment. In addition, AI has the potential to augment the reach of the researcher and therapist by expanding our ability to gather data and deliver interventions beyond the confines of the lab or clinical office.

Reports on remote psychotherapies for youth (e.g., technology-based treatment) suggest it is acceptable, feasible, and useful in overcoming logistical barriers to treatment. But how effective is remote care? To find out, PsycINFO and PubMed were searched from 1960 through 2020, supplemented by journal searches and reference trails, to identify randomized controlled trials of youth psychotherapy for anxiety (including obsessive–compulsive disorder and trauma), depression, attention-deficit/hyperactivity disorder (ADHD), or conduct problems, in which all therapeutic contact occurred remotely. Articles (N = 37) published from 1988 through 2020, reporting 43 treatment-control group comparisons, were identified. Robust variance estimation was used to account for effect size dependencies and to synthesize overall effects and test candidate moderators. Pooled effect size was .47 (95% confidence interval [CI: .26, .67], p < .001) at posttreatment, .44 (95% CI [.12, .76], p < .05) at follow-up—comparable to effects reported in meta-analyses of in-person youth psychotherapy. Effects were significantly (a) larger for remote psychotherapies supported by therapeutic provider contact (.64) than for those accessed by youths, with only logistical support (.22), (b) larger for treatments with phone contact (.65) than for those without (.25), (c) larger for treatment of anxiety (.62) and conduct problems (.78) than ADHD (–.03), and (d) smaller for therapies involving attention/working memory training (–.18) than for those without (.60). Among studies with therapeutic contact, effects were significantly larger when therapists facilitated skill-building (e.g., practicing exposures or problem solving [.68]) than when therapists did not (.18). These findings support the effectiveness of remote psychotherapies for youths, and they highlight moderators of treatment benefit that warrant attention in future research.

No abstract