CoChat
Sign Up
Cited by 100Digital Mental Health InterventionsArtificial Intelligence in Healthcare and EducationSocial Robot Interaction and HRI

Waiting for a digital therapist: three challenges on the path to psychotherapy delivered by artificial intelligence

Open access
Grodniewicz, J. P., Hohol, Mateusz·2023·Frontiers in Psychiatry
DOI 10.3389/fpsyt.2023.1190084PMC PMC10267322
Open this paper in CoChat — free accountOpen PDF

Abstract

Growing demand for broadly accessible mental health care, together with the rapid development of new technologies, trigger discussions about the feasibility of psychotherapeutic interventions based on interactions with Conversational Artificial Intelligence (CAI). Many authors argue that while currently available CAI can be a useful supplement for human-delivered psychotherapy, it is not yet capable of delivering fully fledged psychotherapy on its own. The goal of this paper is to investigate what are the most important obstacles on our way to developing CAI systems capable of delivering psychotherapy in the future. To this end, we formulate and discuss three challenges central to this quest. Firstly, we might not be able to develop effective AI-based psychotherapy unless we deepen our understanding of what makes human-delivered psychotherapy effective. Secondly, assuming that it requires building a therapeutic relationship, it is not clear whether psychotherapy can be delivered by non-human agents. Thirdly, conducting psychotherapy might be a problem too complicated for narrow AI, i.e., AI proficient in dealing with only relatively simple and well-delineated tasks. If this is the case, we should not expect CAI to be capable of delivering fully-fledged psychotherapy until the so-called “general” or “human-like” AI is developed. While we believe that all these challenges can ultimately be overcome, we think that being mindful of them is crucial to ensure well-balanced and steady progress on our path to AI-based psychotherapy.

AI snapshot

Written from the paper’s full text.

TL;DR

This conceptual paper identifies three obstacles to fully fledged psychotherapy delivered by conversational AI: uncertainty about psychotherapy’s active ingredients, whether a non-human agent can form a therapeutic relationship, and whether narrow AI can handle psychotherapy’s complexity. It argues these challenges may eventually be overcome but that current chatbots should be treated as supplements, not autonomous therapists.

Key findings

  • The paper formulates three central challenges: The Problem of a Confused Therapist (limited understanding of what makes psychotherapy effective), The Problem of a Non-human Therapist (whether psychotherapy requires a therapeutic relationship), and The Problem of a Narrowly Intelligent Therapist (whether narrow AI can conduct psychotherapy or whether general AI is needed).
  • Existing mental health chatbots are widespread—more than 10,000 mental health apps exist—and a meta-analysis of 32 randomized controlled trials found conversational AI interventions effective, particularly on a proximal time scale, in reducing depressive and anxiety symptoms, preventing stress, general distress, and negative affect, and improving well-being.
  • The authors argue the “efficacy overflow argument” is invalid: evidence that a psychotherapeutic approach works when delivered by a trained human therapist does not mean chatbot-delivered techniques based on that approach are evidence-based psychotherapy.
  • For the non-human therapist problem, they outline three strategies—Deflation, Mimicry, and Emulation—and note that digital therapeutic alliance can be measured with tools such as mARM and D-WAI, though current understanding remains restricted. For narrow AI, they pose a dialogue-system dilemma: general-purpose systems are uncontrollable and unpredictable, while predictable frame-based systems are restricted to narrow tasks.

Methods

  • Design: conceptual/theoretical analysis, not an empirical study; the authors review literature and formulate three challenges on the path to AI-delivered psychotherapy.
  • Data or sample: no human participants or new dataset; illustrative examples include Woebot dialogues, the MYLO chatbot, and digital therapeutic alliance measures.
  • Main analysis: argumentative and philosophical discussion of psychotherapy traditions, common factors, therapeutic relationship, narrow vs. general AI, and dialogue-system design; no statistical analysis is reported.

Limitations

  • Stated: current understanding of the digital therapeutic alliance is restricted and requires further research with novel measurements; existing measures such as mARM and D-WAI may be incomplete because they adapt traditional alliance measures by replacing “therapist” with “app.”
  • Stated: Mimicry raises ethical problems because it is a form of deception, Emulation’s feasibility is difficult to assess, and the Problem of the Narrowly Intelligent Therapist remains open.
  • Evident from the text: as a conceptual paper, it presents no new empirical data, no systematic review protocol, and no outcome measures; examples and cited studies are used illustratively rather than as a formal evidence synthesis.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.3389/fpsyt.2023.1190084

Cites

125 of 125 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 59 cited sources checked · 0 supported · 0 flagged

  • Interpretationcitation

    The lack of transparency regarding the actual nature of services provided by chatbots is probably most visible in the Efficacy Overflow Argument, often implicitly conveyed in the marketing of mental health chatbots.

    “The lack of transparency regarding the actual nature of services provided by chatbots is probably most visible in what we may call “The Efficacy Overflow Argument,” often implicitly conveyed in the marketing of mental health chatbots (29).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [29] Tekin, 2021, Is big data the new stethoscope? Perils of digital phenotyping to address mental illness, Philos Technol source

      Not checked in the free run.

  • Interpretationreasoning

    If conducting psychotherapy cannot be clearly delineated or divided into a set of clearly delineated tasks, we will not be able to develop AI systems conducting fully fledged psychotherapy unless we first build a general AI.

    “If not, it does not seem that we will be able to develop AI systems conducting fully fledged psychotherapy unless we first build a general AI, i.e., AI which – very much like a human – would be able to apply its intelligence to an unrestricted set of tasks.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: In the proximate future, AI systems available for psychotherapy will be narrow AI.
    • Rests on an assumption: Narrow AI systems are very effective but only in performing well-delineated and clearly constrained tasks.
  • Premisenone given

    Companies developing mental health chatbots are increasingly cautious not to characterize the services offered by chatbots as psychotherapy.

    “Companies developing such chatbots are increasingly cautious not to characterize the services offered by chatbots as psychotherapy.”
    Quote in the paperNot a substantial claim
  • Premisecitation

    The closest we get to AI-based psychotherapy these days are interventions delivered by mental health chatbots, many of which are based on relatively simple dialogue systems.

    “The closest we get to AI-based psychotherapy these days are the interventions delivered by mental health chatbots, many of which are based on relatively simple dialogue systems (34).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [34] Laranjo, 2018, Conversational agents in healthcare: a systematic review, J Am Med Inform Assoc source

      Not checked in the free run.

  • Premisecitation

    Since Rosenzweig, different authors have suggested dozens of lists of common factors potentially responsible for making different therapies effective independently of specific techniques.

    “Since the publication of Rosenzweig (64), different authors suggested dozens of lists of common factors potentially responsible for making different therapies effective independently of what specific techniques they involve (65–70).”
    Quote in the paperSubstantial claim0 of 7 sources checked
    • cites [64] Rosenzweig, 1936, Some implicit common factors in diverse methods of psychotherapy, Am J Orthopsychiatry source

      Not checked in the free run.

    • cites [65] Beutler, 2000, Eclectic psychotherapy

      Not checked in the free run.

    • cites [66] Frank, 1991 source

      Not checked in the free run.

    • cites [67] Goldfried, 1980, Toward the delineation of therapeutic change principles, Am Psychol source

      Not checked in the free run.

    • cites [68] Karasu, 1986, The specificity versus nonspecificity dilemma: toward identifying therapeutic change agents, Am J Psychiatry source

      Not checked in the free run.

    • cites [69] Kleinke, 1994

      Not checked in the free run.

    • cites [70] Wampold, 2015 source

      Not checked in the free run.

  • Classified as probably not a substantial claim. 37
  • May combine two claims. 20

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationreasoning

    If conducting psychotherapy is too complicated for narrow AI, we should not expect CAI to be capable of delivering fully-fledged psychotherapy until general or human-like AI is developed.

    “If this is the case, we should not expect CAI to be capable of delivering fully-fledged psychotherapy until the so-called “general” or “human-like” AI is developed.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given

    Different therapeutic traditions conceptualize the process of psychotherapy differently, often disagreeing about what a therapist should do while conducting therapy.

    “The gist of the problem is that different therapeutic traditions conceptualize the process of psychotherapy differently, often disagreeing about what a therapist should do while conducting therapy.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Premisenone given

    It is not obvious to what extent the active ingredients in effective psychotherapy are specific techniques and interventions as opposed to common factors such as a supportive therapeutic relationship.

    “Part of the problem is that it is not obvious to what extent the active ingredients in effective psychotherapy are specific techniques and interventions as opposed to such common factors as, e.g., a supportive therapeutic relationship.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationnone given

    We might not be able to develop effective AI-based psychotherapy unless we deepen our understanding of what makes human-delivered psychotherapy effective.

    “Firstly, we might not be able to develop effective AI-based psychotherapy unless we deepen our understanding of what makes human-delivered psychotherapy effective.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given

    Conducting psychotherapy might be a problem too complicated for narrow AI, i.e., AI proficient in dealing with only relatively simple and well-delineated tasks.

    “Thirdly, conducting psychotherapy might be a problem too complicated for narrow AI, i.e., AI proficient in dealing with only relatively simple and well-delineated tasks.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    Woebot's exemplary conversation mimics many elements of an actual therapy session.

    “Already in this short conversation, Woebot mimics a lot of elements of an actual therapy session.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Premisenone given

    Clinical psychologists have tried to answer the question of what makes psychotherapy effective for the last 100 years, and we still seem to be far away from a fully satisfactory solution.

    “Clinical psychologists have tried to answer this question for the last 100 years, and we still seem to be far away from a fully satisfactory solution.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationreasoning

    It is unclear whether we will be able to develop an artificial system delivering the processes, techniques, and interventions that make psychotherapy effective unless we first better understand which processes, techniques, and interventions make psychotherapy effective.

    “it is unclear whether we will be able to develop an artificial system delivering the processes, techniques, and interventions that make psychotherapy effective unless we first better understand which processes, techniques, and interventions make psychotherapy effective.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Prochaska and Norcross estimate that there are now more than 500 different psychotherapeutic approaches.

    “Prochaska and Norcross (60) estimate that there are now more than 500 different psychotherapeutic approaches.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [60] Prochaska, 2018

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Most psychotherapeutic approaches belong to one of the main psychotherapeutic traditions, such as psychoanalytic/psychodynamic, existential, person-centered, behavioral, or cognitive.

    “Most of these approaches belong to one of the main psychotherapeutic traditions, e.g., psychoanalytic/psychodynamic, existential, person-centered, behavioral, cognitive, etc.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    In a meta-analysis of 32 randomized controlled trials, CAIs proved effective, particularly on a proximal time scale, in reducing depressive and anxiety symptoms, preventing stress, general distress, and negative affect, and improving well-being.

    “in a recent meta-analysis of 32 randomized controlled trials, Yuhao He and colleagues (57) found that CAIs proved to be effective, in particular on a proximal time scale, in reducing depressive and anxiety (both general and specific) symptoms, preventing stress, general distress, and negative affect, and improving well-being.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [57] He, 2023, Conversational agent interventions for mental health problems: systematic review and meta-analysis of randomized controlled trials, J Med Internet Res source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    According to the assumptions underlying MOL, all that clients need to resolve an internal conflict and regain control is for someone to help them talk about the problem at length, in detail and in the present moment, sustaining their attention to it, and to notice disruptions in their speech and behaviour as they describe the problem, such that the client can shift their attention to aspects of it

    “According to the assumptions underlying MOL, all that clients need to be in a position to resolve such a conflict and regain control is for someone to “(1) help them to talk about the problem at length, in detail and in the present moment, thereby sustaining their attention to it, and (2) to notice disruptions in their speech and behaviour as they describe the problem, such that the client can”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning

    The problem of a confused therapist is relevant to achieving fully fledged AI-based psychotherapy but is not specific to it; the same problem arises in training future human therapists.

    “This problem is relevant to the quest of achieving fully fledged AI-based psychotherapy but not specific to it. We face the same problem in the case of training of future therapists in general.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    MOL is based on an all-encompassing psychological theory: Perceptual Control Theory, according to which the most important principle guiding life is control.

    “MOL is based on an all-encompassing psychological theory: Perceptual Control Theory (80, 81), according to which the most important principle guiding life – from the level of basic biological functioning all the way up to mental health and well-being – is control.”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [80] Carey, 2014, A biopsychosocial model based on negative feedback and control, Front Hum Neurosci source

      Not checked in the free run.

    • cites [81] Powers, 1973

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    Most therapists declare that they integrate methods from different therapeutic orientations.

    “According to Prochaska and Norcross (60), most therapists declare that they integrate methods from different therapeutic orientations.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [60] Prochaska, 2018

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisereasoning

    The therapeutic orientation considered most promising for AI-based therapy is CBT.

    “Simultaneously, judging by the state of the field of mental health chatbots, the therapeutic orientation considered to be most promising for AI-based therapy is CBT.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    If we train algorithms on data from successful psychotherapies, we will have to decide how to understand success in psychotherapy.

    “Moreover, assuming that we would like to train our algorithms on data coming from successful psychotherapies, we will have to decide how we want to understand “success” in psychotherapy.”
    Quote in the paperNot a substantial claim
    • The authors state this as an assumption: We would like to train AI algorithms on data coming from successful psychotherapies.

  • Classified as probably not a substantial claim.

    Premisecitation

    Weinberg discusses common factors including the therapeutic relationship, expectations of therapeutic success, confronting the problem, experience of mastery or control, and attribution of therapeutic success or failure to internal rather than external causes.

    “For example, Weinberg (71) discusses such common factors as: (i) the therapeutic relationship between a therapist and a client/patient, (ii) expectations of therapeutic success, (iii) client’s confronting or facing the problem they struggle with, (iv) experience of mastery or control over the problematic issue, and (v) attribution of therapeutic success or failure to internal causes (e.g., changes”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [71] Weinberger, 1995, Common factors aren’t so common: the common factors dilemma, Clin Psychol Sci Pract source

      Not checked in the free run.

  • May combine two claims.

    Interpretationcitation

    Without a guiding theory or system of psychotherapy, clinicians would be vulnerable, directionless creatures bombarded with hundreds of impressions and pieces of information in a single session; this is equally true about artificial therapists.

    “Prochaska and Norcross (60) point out that “Without a guiding theory or system of psychotherapy, clinicians would be vulnerable, directionless creatures bombarded with literally hundreds of impressions and pieces of information in a single session” [p. 4]. This is equally true about artificial therapists.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [60] Prochaska, 2018

      Not checked in the free run.

  • May combine two claims.

    Interpretationcitation

    In developing artificial therapists, we will have to make the decision about therapeutic orientation for them, either by hard-wiring them with principles of a given therapeutic modality or by training AI algorithms on data from sessions in which a particular therapeutic modality is implemented.

    “But in the case of developing artificial therapists, we will have to make this decision for them, either by hard-wiring them with principles of a given therapeutic modality or by training AI algorithms on data from sessions in which a particular therapeutic modality is implemented (75).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [75] Blackwell, 2020

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Interpretationcitation

    Integrative approaches may prove useful in future attempts to develop new and improve existing therapeutic interventions delivered by CAIs.

    “On the other hand, integrative approaches, “characterized by dissatisfaction with single-school approaches and a concomitant desire to look across school boundaries to see what can be learned from other ways of conducting psychotherapy” in order “to enhance the efficacy, efficiency, and applicability of psychotherapy” (77) [p. 4], may prove useful in our future attempts to develop new, and improve”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [77] Norcross, 2005, A primer on psychotherapy integration source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    Psychological distress experienced by people seeking psychotherapeutic help results from the emergence of an internal conflict, which triggers a loss of control.

    “Psychological distress experienced by people seeking psychotherapeutic help results from the emergence of an internal conflict, which triggers a loss of control.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Premisenone given

    Mimicry is a form of deception, and it is more effective the more its receivers are deceived.

    “In short, mimicry is a form of deception, and it is more effective the more its receivers are deceived.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Some helpful psychological interventions do not require a therapeutic relationship; for example, various writing techniques are effective in reducing psychological distress in certain populations.

    “One more thing worth keeping in mind in this context is that we already know some helpful psychological interventions that do not require a therapeutic relationship. A notable example is various writing techniques, which turn out – at least in certain populations – to be effective in reducing psychological distress (99).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [99] Allen, 2020, Online writing about positive life experiences reduces depression and perceived stress reactivity in socially inhibited individuals, Psychiatry Res source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given

    Mimicry seems to be the default strategy in modern chatbot development.

    “Mimicry seems to be the default strategy in modern chatbot development.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    The digital therapeutic alliance (DTA) is a user-perceived alliance composed of a bond, agreement on tasks, and agreement on goals, which can occur in interaction with CAI.

    “Thus, following Bordin (101), Kaveladze and Schueller (102) define the digital therapeutic alliance (DTA) as “a user-perceived alliance (composed of a bond, agreement on the tasks directed toward improvement, and agreement on therapeutic goals)” [emphasis added, pp. 88–89], which, at least to some extent, can occur in interaction with CAI.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [102] Kaveladze, 2023, A digital therapeutic alliance in digital mental health source

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    Existing mental health chatbots are not persons and probably not even agents, given their restricted and limited capacity to act.

    “Existing mental health chatbots are obviously not persons, and probably not even agents, given how restricted and limited their capacity to act.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Premisecitation

    Psychotherapy is more than the delivery of specific techniques; it is first and foremost an interpersonal relationship.

    “According to most authors, psychotherapy is much more than the delivery of specific techniques. It is first and foremost an interpersonal relationship (30, 45, 94, 95).”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites [30] Tekin, 2023, Ethical issues surrounding artificial intelligence technologies in mental health: psychotherapy chatbots

      Not checked in the free run.

    • cites [45] Manriquez Roa, 2021, The ethics of artificial intelligence in psychotherapy source

      Not checked in the free run.

    • cites [94] Askjer, 2021, The working alliance in blended versus face-to-face cognitive therapy for depression: a secondary analysis of a randomized controlled trial, Internet Interv source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Premisecitation

    The basic idea of CBT - that self-relevant thoughts, evaluations, and beliefs are key contributors to the development and persistence of psychopathological states - is simple and elegant.

    “Moreover, the basic idea of CBT – that “self-relevant thoughts, evaluations, and beliefs are key contributors to the development and persistence of psychopathological states” (42) [p. 23] – is simple and elegant.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [42] Clark, 2014, Cognitive restructuring

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    Therapeutic alliance consists of agreement on therapeutic goals, assignment of therapeutic tasks, and development of bonds.

    “According to a classical characterization offered by Edward Bordin (101), therapeutic alliance consists of (a) an agreement on therapeutic goals, (b) an assignment of therapeutic tasks, and (c) the development of bonds.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [101] Bordin, 1979, The generalizability of the psychoanalytic concept of the working alliance, Psychother Theory Res Pract source

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    A therapeutic relationship is characterized by warmth, accurate empathy, and genuineness.

    “Such a relationship is characterized by “warmth, accurate empathy, and genuineness” (90) [p. 45].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [90] Beck, 1979

      Not checked in the free run.

  • May combine two claims.

    Interpretationcitation

    Existing chatbots allow users to establish DTA, including bonding and agreeing on tasks and goals, which may contribute to reducing symptoms.

    “Thus, we can conclude that existing chatbots allow users to establish DTA, including bonding and agreeing on tasks and goals, which may contribute to reducing symptoms.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    With a global median of 13 mental health workers per 100,000 population (62.2 in high-income countries) and an estimated 70% of people with mental illness receiving no treatment from health care staff, the scenario of everyone having easy access to a well-trained human therapist is unrealistic.

    “Unfortunately, with a global median of 13 mental health workers per 100,000 population (62.2 in high-income countries) (97), and the estimate of 70% of people with mental illness receiving no treatment from health care staff (98), this scenario is unrealistic.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [97] 2021

      Not checked in the free run.

    • cites [98] Henderson, 2013, Mental illness stigma, help seeking, and public health programs, Am J Public Health source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    Clients are autonomously motivated when they experience themselves as having freely chosen their goals and the choice is felt to emanate from themselves.

    “According to Zuroff et al. (127), clients are “autonomously motivated when they experience themselves as having freely chosen their goals and the choice is felt to emanate from themselves” [p. 137].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [127] Zuroff, 2007, Autonomous motivation for therapy: a new common factor in brief treatments for depression, Psychother Res source

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    If we focus only on the behavioral aspect of empathy in developing artificial agents, we may be developing more sophisticated forms of mimicry, and thus the emulation strategy collapses to the mimicry strategy.

    “But if, while developing artificial agents, we focus only on the behavioral aspect of empathy, neglecting all other aspects, such as emotional and phenomenal, maybe we are just developing more sophisticated forms of mimicry, and thus strategy three collapses to strategy two.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Premisenone given

    Human mental suffering comes in a myriad of different forms.

    “Just as unhappiness of unhappy families, human mental suffering comes in a myriad of different forms.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Premisenone given

    We are still waiting for artificial general intelligence (AGI), an artificial system capable of applying its intelligence to a virtually unrestricted range of tasks and environments, including new ones.

    “Even though narrow artificial intelligence is gradually becoming “less narrow,” we are still waiting for the so called artificial general intelligence (AGI), i.e., an artificial system capable of applying its intelligence to a virtually unrestricted range of tasks and environments, including ones that are new to it.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    As long as we want ANIs to primarily implement specific therapeutic techniques, we must keep the range of topics and tasks they can engage in strictly restricted.

    “However, as long as we would like our ANIs to primarily implement specific therapeutic techniques, we would also have to keep the range of the topics and tasks they can engage in strictly restricted.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Mental health chatbots can be very helpful to at least some of their users.

    “While the available research demonstrates that mental health chatbots can, undoubtedly, be very helpful to at least some of their users, the goal of this paper was to outline the scope of the challenge of developing fully fledged AI-based psychotherapy.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Interpretationreasoning

    In light of the Problem of the Narrowly Intelligent Therapist, we may be forced to admit either that artificial therapists are impossible until AGI or that 'psychotherapy' will mean something different in the future.

    “In light of it, we may be forced to admit either that artificial therapists are impossible to construe until we reach the level of technological engagement equal to AGI or that “psychotherapy” in the future will mean something different than it means today.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    If an agent capable of conducting fully fledged psychotherapy must be able to engage in a meaningful conversation about all forms of human mental suffering, then an ANI probably cannot do it.

    “Does this mean that an agent capable of conducting fully fledged psychotherapy would have to be able to engage in a meaningful conversation about all of them? If yes, we probably should not expect an ANI to be able to do it.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationcitation

    Whenever there is a place for unscripted natural language input, a chatbot must classify it into coarse-grained, pre-defined categories, which severely limits its ability to appropriately react to the user's input.

    “Whenever there is a place for unscripted, natural language input from the user, a chatbot has to classify it as falling into one of the coarse-grained, pre-defined categories.8 It severely limits the chatbot’s ability to appropriately react to the user’s input, as witnessed, for example, by the following exchange:”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Interpretationcitation

    If autonomy is the common factor, CAIs may be well positioned to emulate it by creating a context in which clients work through their problems without feeling dependent, paternalized, or forced by an over-imposing therapist.

    “If this is the case, CAIs may be well positioned to emulate it, simply by creating a context in which clients work through their problems without the impression of being dependent on, paternalized by, or forced into something they did not choose by an over-imposing therapist.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    Autonomy is the common factor of most efficient therapeutic interventions, and the main task of a therapist is to support rather than undermine it.

    “Zuroff and colleagues suggest that such understood autonomy is the common factor of most efficient therapeutic interventions, and the main task of a therapist is to support rather than undermine it (128).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [128] Markland, 2005, Motivational interviewing and self–determination theory, J Soc Clin Psychol source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given

    At this point, we are stuck with ANI.

    “Be it as it may, at this point, we are stuck with ANI, and thus the last problem on the way towards AI-based psychotherapy is the following:”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Premisecitation

    Intelligence measures an agent's ability to achieve goals in a wide range of environments.

    “Based on the analysis of multiple available definitions, Shane Legg and Marcus Hutter (129) propose the following working definition of intelligence for AI research: “Intelligence measures an agent’s ability to achieve goals in a wide range of environments” [p. 402].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [129] Legg, 2007

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationnone given

    Human therapists possess many (if not most) skills necessary to conduct psychotherapy just by virtue of successfully participating in everyday social interactions.

    “One benefit of posing this problem is that it sheds light on an issue which we typically neglect in the context of training human therapists, namely, that they possess many (if not most) skills necessary to conduct psychotherapy just by virtue of successfully participating in everyday social interactions.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Premisereasoning

    The artificial intelligence available now can achieve goals only in a relatively constrained or narrow range of environments, and is therefore called artificial narrow intelligence (ANI).

    “The kind of artificial intelligence available now can achieve goals only in a relatively constrained or “narrow” range of environments (or solve only a constrained and “narrow” range of tasks), therefore it is typically called artificial narrow intelligence (ANI).”
    Quote in the paperSubstantial claim
48 of 48

  • The authors state this as an assumption.

    Psychotherapy requires building a therapeutic relationship.

    “assuming that it requires building a therapeutic relationship, it is not clear whether psychotherapy can be delivered by non-human agents.”

  • The paper relies on this without support.

    In the proximate future, AI systems available for psychotherapy will be narrow AI.

    “Despite the hype and enthusiasm around AI research, at least in the proximate future, we are stuck with narrow AI technologies.”

  • The paper relies on this without support.

    Narrow AI systems are very effective but only in performing well-delineated and clearly constrained tasks.

    “Narrow (as opposed to general) AI systems are very effective but only in performing well-delineated and clearly constrained tasks, such as image recognition, speech processing and generation, or even driving a car.”

  • The authors state this as an assumption.

    We would like to train AI algorithms on data coming from successful psychotherapies.

    “assuming that we would like to train our algorithms on data coming from successful psychotherapies, we will have to decide how we want to understand “success” in psychotherapy.”
4 of 4

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Author name differs from the cited reference

    Minor

    The text attributes reference 71 to Weinberg, but the reference list names its author Weinberger.

  • Access date precedes the cited publication year

    Minor

    Reference 85 is dated 2023 but says it was accessed in 2022.

2 of 2

Cites

1 retracted

Flagged first, then by citations

Research works printed but not linked: 6

  1. Tekin, 2023, Ethical issues surrounding artificial intelligence technologies in mental health: psychotherapy chatbots
  2. Norcross, 1990, An eclectic definition of psychotherapy
  3. Kosinski, 2023, Theory of mind may have spontaneously emerged in large language models
  4. Beutler, 2000, Eclectic psychotherapy, 128
  5. Beck, 1979, Cognitive therapy of depression
  6. Schlosser, 2019, Agency

Cited by

Sorted by citations

Related

Authors
Grodniewicz, J. P., & Hohol, M.
Affiliations
Jagiellonian University (PL)
Corresponding author
  • Grodniewicz, J. P.
  • Hohol, Mateusz
ORCID
Venue
Frontiers in Psychiatry
Publisher
Frontiers Media SA
Published
2023-06-01
Volume
14
Pages
1190084-1190084
ISSN
1664-0640 (ISSN-L)
DOI
10.3389/fpsyt.2023.1190084
PMC
PMC10267322
OpenAlex
W4379010346
Semantic Scholar
c0bd3bffc21578ad487fb8b2b2dab704afa5902e
Cited by
100
References
122
Citations by year
2026: 30 · 2025: 40 · 2024: 24 · 2023: 6
FWCI
25.51
Type
journal article
Language
en
Open access
Licence
Locations
  • Frontiers in Psychiatry (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Homo Politicus (Academy of Humanities and Economics in Lodz) (submittedVersion, cc-by-nc-nd)
  • PhilPapers (PhilPapers Foundation) (submittedVersion, other-oa)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • PhilPapers (PhilPapers Foundation) (submittedVersion)
Funders
National Science Centre, Poland (2021/43/B/HS1/02868)
Topics
  • Social Sciences › Psychology › Applied Psychology › Digital Mental Health Interventions
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
  • Social Sciences › Psychology › Social Psychology › Social Robot Interaction and HRI
Keywords
conversational AI, therapeutic relationship, artificial general intelligence, narrow AI, access to mental health care, chatbots

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-07.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-07.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review