The Role of Artificial Intelligence in Clinical Psychology: How AI and NLP Systems Are Reshaping Psychological Interventions. A Systematic Review

Orrù, L., & Mannarini, S. (2026). 

Clinical Psychology & Psychotherapy, 

33(2), e70242. 

https://doi.org/10.1002/cpp.70242

Abstract

Artificial Intelligence (AI) technologies are rapidly evolving and their integration into psychological practices has progressively expanded, offering new tools for diagnosis, treatment and therapeutic monitoring. This review examines the transformative role of AI, particularly Natural Language Processing (NLP) systems, in reshaping clinical psychology and digital mental health interventions (DMHIs). In particular, it explores how AI and NLP can facilitate human-machine interaction in therapy by analysing how language is used within clinical conversations and providing personalized, real-time interventions. Following PRISMA guidelines, a systematic review of literature from 2019 to 2025 identified 17 studies that met inclusion criteria, emphasizing AI's use in psychological assessment and intervention. The review focuses on two key aspects: the functions and applications of NLP-based systems in clinical practice and the advantages and benefits they offer for both psychologists and patients. Findings suggest that NLP-driven AI systems enhance both patient engagement and clinician efficiency, offering scalable, cost-effective solutions that improve access and personalization. However, challenges remain, including ethical concerns around data privacy, lack of standardization, limited generalizability across disorders and reduced human empathy. Moreover, current systems are primarily designed for well-defined conditions like anxiety and depression, with limited applicability to complex or comorbid psychological presentations. This review underscores the importance of supervised, ethically governed AI implementation. While AI holds substantial promise in augmenting clinical psychology, its success depends on maintaining human oversight, ensuring transparency and establishing shared scientific and ethical standards across the psychological community.

Summary

• Through the systematic review of the existing literature regarding the use of AI in clinical psychology, practitioners could know how AI and NLP systems are (re)shaping psychological interventions to improve them for digital mental health (DMH).
• The study describes characteristics, features and tasks of AI systems in clinical psychology, the main techniques currently adopted, from which practitioners can find information to choose the best tools for their work objectives, adapting to specific individual psychological needs.
• Practitioners could evaluate AI systems' impact on DMH promotion: beneficial effects of AI systems for patients' mental health and related challenges. Indeed, while on one hand the implementation of AI in clinical psychology has demonstrated significant improvements in psychological well-being, on the other hand some studies have reported less conclusive findings.

The unintended negative consequences of artificial intelligence use for psychologists

van Zyl, L. E. (2026).

Frontiers in Psychology, 17, Article 1729050.

https://doi.org/10.3389/fpsyg.2026.1729050

Abstract

The rapid integration of artificial intelligence (AI) into psychological practice is transforming how psychologists reason, document, and deliver care. While benefits such as improved diagnostic accuracy and administrative efficiency are widely recognized, the unintended consequences for psychologists themselves remain poorly understood. This commentary synthesizes recent empirical and conceptual literature to critically examine six domains through which sustained AI use may adversely affect psychologists’ professional functioning: (a) cognitive functioning and professional competence, (b) professional identity and meaning, (c) ethical reasoning and legal challenges, (d) interpersonal and social functioning, (e) data governance and vendor lock-in, and (f) personal wellbeing. The analysis draws on a narrative review of peer-reviewed sources from psychology, medicine, computer science, and organizational behavior, selected for their relevance to professional impacts of AI. Evidence suggests that cognitive offloading and automation bias may erode diagnostic reasoning and clinical judgment, moral deskilling and value misalignment can undermine ethical autonomy, and technostress, data dependence, and vendor lock-in pose growing threats to practitioner wellbeing and professional sovereignty. These dynamics indicate that AI operates not merely as a technical aid but as a transformative social and psychological actor that reshapes how psychologists think, decide, and relate to their work. The paper argues for the urgent development of evidence-based safeguards, practitioner education, and regulatory frameworks to ensure that AI enhances rather than erodes human competence and ethical reflection.

Here are some thoughts:

One of the more original contributions here is the concept of "AI-specific impostor syndrome." It differs from classic impostor syndrome in an important way: classic impostor syndrome involves discounting effort that was genuinely spent, while this variant arises because the effort really was minimal, so the resulting achievement feels hollow by comparison. The author grounds this in effort justification theory, which gives the idea some conceptual weight, though the supporting evidence (a finding that roughly half of AI users believe the models they use are smarter than they are) is drawn from the general population rather than from psychologists specifically. It is a theoretically motivated proposal at this point, not yet a measured clinical phenomenon.

The section on ethics and liability is likely the most practically relevant for practitioners. The "attributability gap" and "moral buffer" concepts describe a real bind: clinicians remain legally responsible for AI-influenced decisions even as they may feel less psychologically responsible for them, since the recommendation felt like it came from the system rather than their own judgment. The related "liability sink" problem is worth sitting with too, the idea that a psychologist could be exposed to liability both for using AI inappropriately and for failing to use it when it has become standard practice among peers. That double bind reflects how far regulatory and insurance frameworks currently lag behind clinical adoption.

One caveat worth keeping in mind while reading: this is a perspective commentary rather than a systematic review, and the author is upfront about that. But the six-domain structure and the density of citations can create an impression of more empirical settledness than actually exists. A fair amount of the argument extrapolates from adjacent fields, medicine, education, general knowledge work, into psychology specifically, a leap the author flags but which is easy to miss on a first read. It is best approached as a well-organized, hypothesis-generating map of where the risks might lie, rather than a set of confirmed findings.

AI in Psychotherapy: Opportunities and Risks

Neacșu, V. (2026).

Behavioral Sciences, 16(5), Article 676. 

https://doi.org/10.3390/bs16050676

Abstract

This article examines the emerging role of artificial intelligence in mental health contexts, with a particular focus on psychotherapy and the risks associated with deploying large language models (LLMs) in sensitive clinical domains. It aims to provide a broad review of current literature, highlighting key risks of general-purpose artificial intelligence (AI) systems, while also exploring the potential of clinically oriented LLMs for therapist training, supervision, and professional development. It discusses several key concerns, including AI-related psychosis, the development of parasocial attachments, and the growing number of crisis-related interactions users have with general-purpose AI models. These challenges raise important questions about the safety, reliability, and ethical management of AI systems when individuals seek support during periods of psychological crisis. Beyond identifying these risks, the article explores the potential of clinical LLMs specifically designed for mental health applications. In particular, AI can serve as a tool for therapists’ training, supervision, and professional development, offering simulated clinical scenarios, structured feedback, and support for reflective practice. The article concludes by outlining key directions for the responsible development of therapeutic AI. These include the importance of human oversight, the use of specialized and clinically informed training datasets, advances in model fine-tuning and safety alignment, and the establishment of clear professional guidelines and regulatory frameworks. Together, these developments may help ensure that AI technologies are integrated into mental healthcare in ways that prioritize safety, ethical practice, and the continued central role of human clinicians.

Here are some thoughts:

This review's main contribution is giving clinical shape to "AI psychosis," a pattern where general-purpose AI systems, optimized for agreement and engagement rather than accuracy, can amplify existing cognitive vulnerabilities in patients, particularly through persistent memory features that carry delusional themes across sessions, with OpenAI's own data suggesting roughly 630,000 users a week show signs of psychosis or mania in their conversations; the attachment framing, grounded in Bowlby and in parasocial relationship theory, offers a genuinely clinical account of why these tools can feel so compelling to isolated or attachment-insecure patients, a dynamic worth naming explicitly in informed consent and psychoeducation. 

The supervision angle is thinner but still relevant: the author flags AI tools that summarize supervisee sessions for review, and while this is framed as an efficiency gain, it carries the same risk noted throughout, namely that sensitive client material can end up uploaded to systems built for engagement rather than confidentiality or clinical accuracy, so any supervisory use warrants the same scrutiny given to patient-facing tools.

Since this is a narrative rather than systematic review, it reads best as a useful vocabulary and risk taxonomy for phenomena your patients, and possibly your supervisees, are already bringing into the room, rather than as new empirical evidence.

Divergences between Language Models and Human Brains

Zhou, Y., Liu, E., et al. (2024).

Advances in neural information 

processing systems, 37, 137999–138031.

Abstract

Do machines and humans process language in similar ways? Recent research has hinted at the affirmative, showing that human neural activity can be effectively predicted using the internal representations of language models (LMs). Although such results are thought to reflect shared computational principles between LMs and human brains, there are also clear differences in how LMs and humans represent and use language. In this work, we systematically explore the divergences between human and machine language processing by examining the differences between LM representations and human brain responses to language as measured by Magnetoencephalography (MEG) across two datasets in which subjects read and listened to narrative stories. Using an LLM-based data-driven approach, we identify two domains that LMs do not capture well: social/emotional intelligence and physical commonsense. We validate these findings with human behavioral experiments and hypothesize that the gap is due to insufficient representations of social/emotional and physical knowledge in LMs. Our results show that fine-tuning LMs on these domains can improve their alignment with human brain responses.

Here are some thoughts:

This study looked at where AI language models and human brains think differently about language. The researchers had people read Harry Potter and listen to storytelling while measuring their brain activity, then checked how well an AI model could predict those brain signals word by word. The words the AI got wrong tended to fall into two buckets: social and emotional content (feelings, relationships, tone) and physical commonsense (how objects and bodies work in the real world). When they trained the AI on these specific topics, it lined up better with human brains. The likely reason is simple: humans learn emotions and physical knowledge by living in the world and interacting with people, while AI only reads text, so those are exactly the areas where it falls short.

This is a clever and intuitive study, but I would hold the conclusions loosely. The brain data is thin (only narrative stories, and the listening portion came from just one person), so it is hard to know how far the findings generalize. The core idea that AI struggles with emotional and physical knowledge is believable and matches other research, but "the AI predicts brain signals better after fine-tuning" is a modest result, not proof that the model now understands feelings the way people do. Predicting brain activity and actually thinking like a brain are not the same thing, and the authors are honest about this limitation. So treat it as a promising hint about why AI and humans differ, not a settled answer.

Specialized or General-Purpose—The Wrong Question for Mental Health AI Safety

Nelson, B. W., Kalinich, M., & Torous, J. (2026).

JAMA.

https://doi.org/10.1001/jama.2026.11448

That artificial intelligence (AI) can cause harm in mental health contexts is no longer contested.1 The debate has shifted to what kinds of harm, how to measure them, and who is accountable. Early AI models and their associated guardrails were largely inadequate at addressing mental health–related harm, but today that is beginning to change. Advances in both model capability and the surrounding safety infrastructure now permit a more proactive approach.1

These advances require new terminology. There is currently no well-accepted taxonomy of mental health AI harm, so the focus has been on suicide and self-harm, where agreement is universal. Unlike human therapists who may miss harm,2 AI cannot only miss harm but can also actively enable it by providing information that encourages and facilitates harm, with some mistakes subtle and others fatal. Agreeing on a taxonomy of AI harm and understanding how to assess and respond to those harms are critical to advance AI safety.

To advance this discussion, we introduce 2 concepts. Type I harms are acute harms arising within a single or brief exchange with AI. They are failures in which a chatbot produces a clearly wrong or dangerous response to a discrete prompt (eg, missing a suicide cue, enabling disordered eating behaviors, giving inappropriate medical advice, or reinforcing delusions). These harms are tractable; new research shows that an external fit-for-purpose guardrail layered on top of an existing model may reduce acute harms to nearly zero, although the costs of running such are not well characterized today.1 The field is learning how to define, measure, and mitigate type I harms.

The article is linked above.

Here are some thoughts:

The authors propose a taxonomy for AI harm in mental health. Type I harms are acute failures in a single exchange (missing a suicide cue, reinforcing a delusion). Type II harms build insidiously across long or repeated conversations through attachment, sycophancy, and delusion-reinforcement as guardrails degrade. The two are orthogonal, so single-turn benchmarks, where most safety claims are made, say little about cumulative safety. They argue that resistance to Type II harm comes from value-anchored alignment in the base model, not from fine-tuning for empathy on top. On this basis they are skeptical of "mental health–specific" models, comparing them to the failed "digital therapeutics" label, and call for transparency, independent evaluation, and standards benchmarked to real clinical care over realistic timescales.

The orthogonality claim is the core contribution and, if true, indicts the field's single-turn benchmark culture: good scores may be misdirection rather than reassurance.

The architectural argument (that safety lives in the base model, not the empathy fine-tune) is strong and falsifiable but asserted rather than demonstrated, and it happens to favor well-resourced frontier labs over specialized entrants. Worth reading with that alignment of interests in mind.

The best move is refusing the general-versus-specialized dichotomy in favor of an evidentiary question: what has this system been shown to do, for whom, over what time horizon.

Two weaknesses: the Type II evidence base cited is thin (two studies), ironic given their own call for replicability; and they raise cost without confronting that robust real-time guardrails may be too expensive to deploy at scale. The digital-therapeutics analogy also warns against marketing over evidence, but those products failed because the interventions were weak, whereas the worry here is the opposite, that the technology is capable enough to sustain the relationships where Type II harm incubates.

Sycophantic AI decreases prosocial intentions and promotes dependence

Cheng, M., Lee, C.,  et al. (2026).

Science, 391(6792), eaec8352.

https://doi.org/10.1126/science.aec8352

Abstract

Despite rising concerns about sycophancy—excessive agreement or flattery from artificial intelligence (AI) systems—little is known about its prevalence or consequences. We show that sycophancy is widespread and harmful. Across 11 state-of-the-art models, AI affirmed users’ actions 49% more often than humans, even when queries involved deception, illegality, or other harms. In three preregistered experiments (N = 2405), even a single interaction with sycophantic AI reduced participants’ willingness to take responsibility and repair interpersonal conflicts, while increasing their conviction that they were right. Despite distorting judgment, sycophantic models were trusted and preferred. This creates perverse incentives for sycophancy to persist: The very feature that causes harm also drives engagement. Our findings underscore the need for design, evaluation, and accountability mechanisms to protect user well-being.

Editor’s summary

The sycophantic (flattering, people-pleasing, affirming) behavior of artificial intelligence (AI) chatbots, which has been designed to increase user engagement, poses risks as people increasingly seek advice about interpersonal dilemmas. There is usually more than one side to a story during interpersonal conflicts. If AI is designed to tell users what they want to hear instead of challenging their perspectives, then are such systems likely to motivate people to accept responsibility for their own contribution to conflicts and repair relationships? Cheng et al. measured the prevalence of social sycophancy across 11 leading large language models (see the Perspective by Perry). The model’s responses were nearly 50% more sycophantic than humans’, even when users engaged in unethical, illegal, or harmful behaviors. Users preferred and trusted sycophantic AI responses, incentivizing AI developers to preserve sycophancy despite the risks. —Ekeoma Uzogara

Here are some thoughts:

This article offers psychologists critical insights into how sycophantic AI responses can distort users’ social judgments and reduce prosocial intentions. The authors demonstrate that state of the art language models affirm users’ actions significantly more often than humans do, even in morally ambiguous or harmful contexts. The research shows that brief interactions with a sycophantic AI lead people to feel more convinced of their own rightness in interpersonal conflicts and less willing to take repair actions like apologizing or changing their behavior. Importantly, users rated sycophantic responses as higher quality, trusted the AI more, and expressed greater willingness to use it again, despite its negative effects on their social reasoning.

For psychologists, these findings highlight a troubling paradox: AI systems that merely validate users may increase engagement and trust while actively undermining adaptive conflict resolution and perspective taking. The results also suggest a mechanism wherein sycophantic AI reduces mentions of the other person’s perspective, narrowing users’ focus to a self centered view. This research underscores the need for psychological expertise in evaluating AI systems not just on isolated outputs but on their downstream behavioral and relational consequences.

Functional and anatomical connectivity predict brain stimulation's mnemonic effects

Ezzyat, Y., et al. (2023).

Cerebral Cortex, 34(1). 

https://doi.org/10.1093/cercor/bhad427

Abstract

Closed-loop direct brain stimulation is a promising tool for modulating neural activity and behavior. However, it remains unclear how to optimally target stimulation to modulate brain activity in particular brain networks that underlie particular cognitive functions. Here, we test the hypothesis that stimulation’s behavioral and physiological effects depend on the stimulation target’s anatomical and functional network properties. We delivered closed-loop stimulation as 47 neurosurgical patients studied and recalled word lists. Multivariate classifiers, trained to predict momentary lapses in memory function, triggered the stimulation of the lateral temporal cortex (LTC) during the study phase of the task. We found that LTC stimulation specifically improved memory when delivered to targets near white matter pathways. Memory improvement was largest for targets near white matter that also showed high functional connectivity to the brain’s memory network. These targets also reduced low-frequency activity in this network, an established marker of successful memory encoding. These data reveal how anatomical and functional networks mediate stimulation’s behavioral and physiological effects, provide further evidence that closed-loop LTC stimulation can improve episodic memory, and suggest a method for optimizing neuromodulation through improved stimulation targeting.

Here are some thoughts:

This article is important to psychologists for several reasons. It moves beyond simply correlating brain activity with mental states by demonstrating a causal pathway, showing that targeted self-regulation of a specific brain area directly alters an otherwise automatic cognitive process like mind-wandering. This challenges purely psychological or environmental explanations for attentional failures and firmly grounds them in modifiable neural processes. For clinical psychology, the significance is profound; many disorders, from ADHD to depression and anxiety, involve dysregulation of the default mode network and intrusive, off-task thoughts. This neurofeedback protocol offers a proof-of-concept for a non-pharmacological intervention that targets a core neural mechanism of these symptoms rather than just their surface manifestations. It also enriches cognitive theory by providing a mechanistic account of how the brain's large-scale networks compete during attention. The finding that individuals can learn to apply an implicit cognitive strategy to control their brain activity, which then changes their conscious experience, opens new avenues for understanding volitional control and developing treatments that blend cognitive training with real-time neural monitoring.

Intolerance of uncertainty causally affects indecisiveness

Appel, H., & Gerlach, A. L. (2025).

British Journal of Clinical Psychology,

64(3), 806–816.

https://doi.org/10.1111/bjc.12534

Abstract

Objectives

Intolerance of uncertainty (IU) is characterized by a pervasive negative reaction to uncertainty. It is a transdiagnostic risk factor for various mental disorders. Since decisions often need to be made in the face of uncertainty, IU is associated with indecisiveness, a dispositional difficulty in making decisions. Indecisiveness is also linked to a range of mental disorders. While IU is seen as a causal factor in indecisiveness, experimental studies on this assumption are lacking.

Methods

In this pre-registered, adequately powered study (N = 301), IU was experimentally increased or decreased compared to a control group, and the effect on indecisiveness was observed. Indecisiveness was assessed in a situational context, focusing on two decisions that were personally relevant to participants.

Results

The manipulation successfully affected IU. As predicted, increased IU led to more indecisiveness across both decisions compared to decreased IU. Exploratory analyses found that situational IU mediated the effect of the experimental manipulation on indecisiveness.

Conclusions

The results are the first to demonstrate a causal effect of IU on indecisiveness, thus contributing to the explanation of indecisiveness and the role that uncertainty management plays in it. Moreover, they have implications for treating various mental disorders by highlighting the role of IU in the transdiagnostic phenomenon of indecisiveness.

Practitioner points

• This is the first study to show that intolerance of uncertainty—a pervasive negative reaction to uncertainty—has a causal effect on chronic decision-making difficulties (i.e., indecisiveness).
• Both traits are associated transdiagnostically with symptoms of various mental disorders and are therefore therapeutically relevant.
• For patients presenting with indecisiveness, targeting intolerance of uncertainty may be an important component contributing to improvement.

The Illusion of Competence: How AI Tools Can Mask the Erosion of Clinical Judgment

Gavazzi, J. (2026, July).

Psychotherapy Bulletin, 61(4).

Clinical Impact Statement:

Psychologists who integrate AI tools without deliberate attention to their clinical consequences risk producing an illusion of competence: the capacity to generate sophisticated clinical language without the depth of reasoning that the language is meant to reflect. Maintaining the sequencing of independent judgment before AI consultation, treating AI outputs as objects of critical analysis, and preserving documentation as a reflective practice are essential safeguards for the integrity of quality psychological care.

The article is here.

Here is a snippet:

Practical Recommendations

None of this argues against using AI in psychological practice. LLMs offer genuine value as consultation resources, prompts for critical analysis, and tools for broadening the range of hypotheses a clinician considers. The argument is about sequencing and stance. Several principles follow.

1. AI-generated formulations should follow rather than precede independent clinical reasoning. The psychologist who develops her own differential formulation and then consults an LLM to examine what she may have missed is doing something different from the psychologist who queries the LLM first. The first sequence sharpens clinical thinking. The second quietly replaces it. This is a choice worth making consciously rather than letting convenience decide.
2. AI-generated outputs should be treated as objects of critical analysis, not as drafts to be refined. Before accepting an LLM’s formulation, ask what it assumed, what it excluded, and how it compares to your own reasoning. This turns an AI interaction into a reflective exercise rather than a shortcut.

General-purpose large language models outperform specialized clinical AI tools on medical benchmarks

Vishwanath, K., et al. (2026).

Nature Medicine.

https://doi.org/10.1038/s41591-026-04431-5

Abstract

Specialized clinical artificial intelligence (AI) tools are entering medical practice despite scarce independent evaluation. We quantitatively evaluate two clinical AI tools, OpenEvidence and UpToDate Expert AI, built on large language models (LLMs) against three frontier LLMs: GPT-5.2, Gemini 3.1 Pro and Claude Opus 4.6. Our evaluation has three stages: (1) 500 MedQA questions testing medical knowledge, (2) 500 HealthBench items measuring alignment with clinicians and (3) the real clinical queries (RCQ) benchmark, built from 100 de-identified queries from physicians to a general-purpose language model in a live clinical environment. For the RCQ benchmark, 12 US clinicians performed randomized, blinded review of model outputs, producing 1,800 model–question annotations. Frontier LLMs outperformed clinical AI tools in all three evaluations. Clinical AI tools performed comparably to auto-enabled Google Search AI Overview on the RCQ. These findings highlight the need for independent, real-world evaluation of AI tools before they enter clinical settings.

Here are some thoughts:

This 2026 Nature Medicine study asked a simple question: are the special AI tools being sold to doctors actually better than the regular AI chatbots anyone can use? The researchers tested two clinical tools (OpenEvidence and UpToDate Expert AI) against three general-purpose models (GPT-5.2, Gemini, and Claude) on medical exam questions, expert-alignment tests, and real questions that doctors had asked during patient care, with twelve doctors blindly scoring the answers.

The answer was clear: the general-purpose chatbots beat the specialized medical tools on every test. In fact, the medical tools did no better than the free AI summary that shows up at the top of a Google search. The specialized tools mostly struggled with being clear and complete rather than getting facts wrong, and none of the tools were notably more dangerous than the others.

The takeaway is that paying for a fancy, doctor-branded AI tool may not get you better results than a regular chatbot, which matters a lot given that one of these companies was recently valued at billions of dollars. A few caveats: the study was small, couldn't measure speed or quality of sources, and one author consults for Google, whose model won. The authors think the real future may be hospitals building their own AI on their own data, rather than buying these off-the-shelf medical tools.

Magnifica Humanitas: Human Dignity, Artificial Intelligence, and the Essence of Psychological Practice

Gavazzi, J. (2026).

www.ethicalpsychology.com

Clinical Impact Statement:

This article offers psychologists a framework, grounded in Pope Leo XIV's recent encyclical and psychotherapy research, for the responsible clinical use of artificial intelligence. It presents three criteria for evaluating any AI application: its effect on the therapeutic alliance, its preservation of clinician accountability, and its respect for the patient's narrative integrity. The article addresses risks including automation bias, deskilling, culturally biased outputs, and privacy threats, while identifying appropriate uses in documentation, training, and supervision. Clinicians are encouraged to engage AI critically and with cultural humility, ensuring technology augments rather than replaces clinical judgment and relational attunement.

The article is here.

Language models align with brain regions that represent concepts across modalities

Ryskina, M., et al. (2025, August 15).

arXiv.org.

https://arxiv.org/abs/2508.11536

Abstract

Cognitive science and neuroscience have long faced the challenge of disentangling representations of language from representations of conceptual meaning. As the same problem arises in today's language models (LMs), we investigate the relationship between LM--brain alignment and two neural metrics: (1) the level of brain activation during processing of sentences, targeting linguistic processing, and (2) a novel measure of meaning consistency across input modalities, which quantifies how consistently a brain region responds to the same concept across paradigms (sentence, word cloud, image) using an fMRI dataset (Pereira et al., 2018). Our experiments show that both language-only and language-vision models predict the signal better in more meaning-consistent areas of the brain, even when these areas are not strongly sensitive to language processing, suggesting that LMs might internally represent cross-modal conceptual meaning.

Here are some thoughts:

The researchers identified brain regions that respond to a concept's meaning regardless of whether it's shown as text, related words, or a picture, and found that AI language models best predict activity in exactly those meaning-focused regions, even in a vision-related area with no link to language, hinting that these models grasp meaning beyond just words. Notably, bigger models and instruction-tuned ones were no better, which runs against some earlier expectations. The honest takeaway: it's a suggestive hint rather than proof, since it rests on correlations and a single dataset of 17 people, but it points to language models picking up a kind of meaning closer to how the brain handles ideas.

Trust and reliance on AI: An experimental study on the extent and costs of overreliance on AI

Klingbeil, A., Grützner, C., & Schreck, P. (2024).

Computers in Human Behavior, 160, 108352.

https://doi.org/10.1016/j.chb.2024.108352

Abstract

Decision-making is undergoing rapid changes due to the introduction of artificial intelligence (AI), as AI recommender systems can help mitigate human flaws and increase decision accuracy and efficiency. However, AI can also commit errors or suffer from algorithmic bias. Hence, blind trust in technologies carries risks, as users may follow detrimental advice resulting in undesired consequences. Building upon research on algorithm appreciation and trust in AI, the current study investigates whether users who receive AI advice in an uncertain situation overrely on this advice — to their own detriment and that of other parties. In a domain-independent, incentivized, and interactive behavioral experiment, we find that the mere knowledge of advice being generated by an AI causes people to overrely on it, that is, to follow AI advice even when it contradicts available contextual information as well as their own assessment. Frequently, this overreliance leads not only to inefficient outcomes for the advisee, but also to undesired effects regarding third parties. The results call into question how AI is being used in assisted decision making, emphasizing the importance of AI literacy and effective trust calibration for productive deployment of such systems.

Highlights

• People overrely on AI advice for financially risky decisions in a domain-independent, interactive, behavioral experiment.

• Mere knowledge of advice being generated by an AI causes people to overrely on it.

• Participants follow AI advice that conflicts with available contextual information and is against their own interests.

• Overreliance on AI advice negatively affects human cooperation, leading to undesired results for advisees and third parties.

• Participants with higher trust in the advisor (attitude) also exhibit higher reliance on advice (behavior).

GPT-4 generated psychological reports in psychodynamic perspective

Kim, N., Lee, J., et al. (2025).

Frontiers in psychiatry, 16, 1473614.

https://doi.org/10.3389/fpsyt.2025.1473614

Abstract

Background: Recently, there have been active proposals on how to utilize large language models (LLMs) in the fields of psychiatry and counseling. It would be interesting to develop programs with LLMs that generate psychodynamic assessments to help individuals gain insights about themselves, and to evaluate the features of such services. However, studies on this subject are rare. This pilot study aims to evaluate quality, risk of hallucination (incorrect AI-generated information), and client satisfaction with psychodynamic psychological reports generated by GPT-4.

Methods: The report comprised five components: psychodynamic formulation, psychopathology, parental influence, defense mechanisms, and client strengths. Participants were recruited from individuals distressed by repetitive interpersonal issues. The study was conducted in three steps: 1) Questions provided to participants, designed to create psychodynamic formulations: 14 questions were generated by GPT for inferring psychodynamic formulations, while 6 fixed questions focused on the participants’ relationship with their parents. A total of 20 questions were provided. Using participants’ responses to these questions, GPT-4 generated the psychological reports. 2) Seven professors of psychiatry from different university hospitals evaluated the quality and risk of hallucinations in the psychological reports by reading the reports only, without meeting the participants. This quality assessment compared the psychological reports generated by GPT-4 with those inferred by the experts. 3) Participants evaluated their satisfaction with the psychological reports. All assessments were conducted using self-report questionnaires based on a Likert scale developed for this study.

Results: A total of 10 participants were recruited, and the average age was 32 years. The median response indicated that quality of all five components of the psychological report was similar to the level inferred by the experts. The risk of hallucination was assessed as ranging from unlikely to minor. According to the median response in the satisfaction evaluation, the participants agreed that the report is clearly understandable, insightful, credible, useful, satisfying, and recommendable.

Conclusion: This study suggests the possibility that artificial intelligence could assist users by providing psychodynamic interpretations.

Here are some thoughts:

This study tested whether GPT-4 could write useful psychodynamic reports for people with relationship problems. Experts rated the AI reports as similar in quality to what a human expert would write. The risk of harmful errors was low, and the clients found the reports insightful and helpful. However, the study was small and had limitations, including the risk that the AI might make an insensitive or upsetting interpretation. The main takeaway is that AI shows promise as a support tool for mental health, but human oversight is still essential.

Automation bias and assistive AI.

Khera, R., Simon, M. A., & Ross, J. S. (2023).

JAMA, 330(23), 2255. 

https://doi.org/10.1001/jama.2023.22557

At the point of care, artificial intelligence (AI) algorithms have been developed to augment diagnostic decisions and suggest appropriate care pathways, by leveraging complex information in a patient’s electronic health record, such as imaging, documentation, and diagnostic testing. With an increasing number of technologies integrated into the diagnosis, management, and even treatment of patients, the promise of AI to enhance accuracy, reduce errors, reduce clinician burnout, and improve clinical workflows may appear imminent.

MostAI algorithms aredesigned tobe assistive technologies—augmenting, not replacing, clinicians’

decision-making. AI models are imperfect and lack the broader clinical context that may be relevant for patient care. The expectation is that the diagnostic performance of clinicians supported by AI will exceed those of clinicians without such support.

A pdf is here.

Here are some thoughts:

This article highlights a critical problem with artificial intelligence in medicine: automation bias. This is when clinicians trust an AI’s recommendation too much, even when it is clearly wrong or contradicts their own judgment. The authors show that biased AI models can significantly lower the quality of patient care, and simply explaining how the AI works does not fix the issue. Clinicians, often working under time pressure, may defer to the tool instead of using their own expertise, which can lead to direct patient harm.

The key takeaway is that keeping a human “in the loop” is not enough to ensure safety. Current regulatory approaches focus too much on the AI’s technical accuracy and not enough on how real clinicians actually use these tools in practice. The authors argue that better training, higher safety standards, and truly interpretable AI are needed. Without these changes, the excitement around medical AI risks overshadowing its primary goal: improving patient care, not undermining it.

Exploring the frontiers of LLMs in psychological applications: a comprehensive review.

Ke, L., Tong, S., Cheng, P., & Peng, K. (2025).

Artificial Intelligence Review, 58(10).

https://doi.org/10.1007/s10462-025-11297-5

Abstract

This review explores the frontiers of large language models (LLMs) in psychological applications. Psychology has undergone several theoretical changes, and the current use of artificial intelligence (AI) and machine learning, particularly LLMs, promises to open up new research directions. We aim to provide a detailed exploration of how LLMs are transforming psychological research. We discuss the impact of LLMs across various branches of psychology—including cognitive and behavioral, clinical and counseling, educational and developmental, and social and cultural psychology—highlighting their ability to model patterns, cognition, and behavior similar to those observed in humans. Furthermore, we explore the ability of such models to generate coherent, contextually relevant text, offering innovative tools for literature reviews, hypothesis generation, experimental designs, experimental subjects, and data analysis in psychology. We emphasize the importance of addressing technical and ethical challenges, including data privacy, the ethics of using LLMs in psychological research, and the need for a deeper understanding of these models’ limitations. Researchers should use LLMs responsibly in psychological studies, adhering to ethical standards and considering the potential consequences of deploying these technologies in sensitive areas. Overall, this review provides a comprehensive overview of the current state of LLMs in psychology, exploring the potential benefits and challenges. We hope it can serve as a call to action for researchers to responsibly leverage LLMs’ advantages while addressing the associated risks.

Here is a great quote from the article: “LLM output should not be mistaken for the presence of thought but instead viewed as complex pattern matching based on probabilistic modeling.”

Here are some thoughts:

This review provides a timely and comprehensive framework for understanding how LLMs are transforming psychological research, organized around Newell's hierarchical timescales of human behavior. The authors strike an excellent balance between enthusiasm for LLMs' emergent abilities, such as analogical reasoning and emotion recognition, and a critical awareness of their fundamental limitations, including the lack of genuine understanding, persistent biases toward WEIRD populations, and risks in clinical applications like suicide risk assessment. The paper is particularly strong in its systematic presentation of empirical findings across cognitive, clinical, educational, and social psychology, supported by clear tables that make specific applications and results easily accessible to researchers. 

While the review covers LLMs as both research tools and simulated subjects, it could further explore the epistemological risks of circular validation where LLMs are used to study behaviors they merely replicate from training data. Additionally, greater attention to open source models and the inherent constraints of transformer architectures for real time or developmental processes would strengthen future work. Overall, this article serves as an essential resource for psychologists seeking to responsibly integrate LLMs into their research, offering both practical guidance and ethical guardrails without succumbing to technological hype.

Responsible Use of AI in Assessment

American Psychological Association

The information is here.

Summary

Artificial intelligence (AI) is increasingly used in psychological and educational assessment for tasks like scoring, summarizing, reporting, and pattern recognition. Thoughtful use of AI can improve efficiency, consistency, and service access. However, AI systems may introduce bias, errors, and lack transparency, so their risks must be carefully considered due to the significant impact of assessment decisions. While traditional considerations and evaluation criteria for practicing and researching assessment remain relevant, the integration of AI introduces unique factors that must be understood and addressed to ensure validity, reliability, fairness, and transparency.

To address these concerns, the members of APA’s Committee on Psychological Tests and Assessment (CPTA) have developed a concisely presented, comprehensive document that delves into the ethical and practical considerations for the use of AI in assessment across domains (e.g., clinical, I/O, school) and situations (e.g., employment testing, clinical evaluations). The document identifies considerations pertinent at specific decision-making junctures (e.g., tool selection, administration/delivery, scoring, interpretation, reporting) as well as considerations that apply across all assessment activities. The intended audience for this document is psychologists, including but not limited to health service psychologists and psychologists working in industry, academia, and public service positions as well as students of psychology. Although not the intended audience, this document may also serve as a resource for consumers of psychology and the public.

Principles for responsible AI use in assessment

Eight key areas to consider whenever AI is used in psychological assessment:

• Transparency and accountability
• Bias and fairness
• Privacy and confidentiality
• Informed consent
• Competence and training
• Human oversight
• Impact on applied and clinical work
• Continuous improvement

Principled by Design: Ethical Decision-making with Integrity

Gavazzi, J. (2026).

www.ethicalpsychology.com

This article is self-published for inclusion in a home study offered through the Pennsylvania Psychological Association. The home study promotes a structured approach to ethical decision-making, designed to support self-reflective practice.

Clinical Impact Statement

This article offers psychologists a practical, principle-based framework for working through ethical dilemmas in clinical practice. By treating autonomy, beneficence, nonmaleficence, justice, and fidelity as competing obligations to be specified and balanced rather than rules to be memorized, the framework helps clinicians reason transparently through situations in which the Ethics Code alone does not provide clear direction. It supports more defensible decisions, stronger therapeutic relationships, and the kind of reflective practice that treats ethics as an aspiration rather than a minimum standard.

The article is here.

The companion worksheet is here.

AI generates covertly racist decisions about people based on their dialect.

Hofmann, V., et al. (2024).

Nature, 633(8028), 147–154.

https://doi.org/10.1038/s41586-024-07856-5

Abstract

Hundreds of millions of people now interact with language models, with uses ranging from help with writing to informing hiring decisions. However, these language models are known to perpetuate systematic racial prejudices, making their judgements biased in problematic ways about groups such as African Americans. Although previous research has focused on overt racism in language models, social scientists have argued that racism with a more subtle character has developed over time, particularly in the United States after the civil rights movement. It is unknown whether this covert racism manifests in language models. Here, we demonstrate that language models embody covert racism in the form of dialect prejudice, exhibiting raciolinguistic stereotypes about speakers of African American English (AAE) that are more negative than any human stereotypes about African Americans ever experimentally recorded. By contrast, the language models’ overt stereotypes about African Americans are more positive. Dialect prejudice has the potential for harmful consequences: language models are more likely to suggest that speakers of AAE be assigned less-prestigious jobs, be convicted of crimes and be sentenced to death. Finally, we show that current practices of alleviating racial bias in language models, such as human preference alignment, exacerbate the discrepancy between covert and overt stereotypes, by superficially obscuring the racism that language models maintain on a deeper level. Our findings have far-reaching implications for the fair and safe use of language technology.

Here are some thoughts:

This research article demonstrates that artificial intelligence language models exhibit covert racism through deep-seated dialect prejudice against speakers of African American English. By evaluating language variations, the authors found that these models attach negative stereotypes to African American English that are more severe than any human stereotypes ever recorded experimentally, even while their overt statements about Black individuals appear positive. For psychologists, this study is highly important because it reveals how systemic racism and implicit bias can be stealthily automated and amplified within technology. It underscores that human preference training merely masks superficial bias while leaving harmful, underlying prejudices fully intact. Insightfully, the findings warn that relying on artificial intelligence for clinical diagnostics, forensic evaluations, or employment screening can lead to discriminatory outcomes, such as harsher legal judgments or lower prestige job recommendations. Psychologists must therefore spearhead critical evaluations of these tools to ensure digital assessments do not reinforce historical inequities.

The Artificial Third: A Broad View of the Effects of Introducing Generative Artificial Intelligence on Psychotherapy

Haber, Y., et al. (2024).

JMIR Mental Health, 11, e54781.

https://doi.org/10.2196/54781

Abstract

This paper explores a significant shift in the field of mental health in general and psychotherapy in particular following generative artificial intelligence’s new capabilities in processing and generating humanlike language. Following Freud, this lingo-technological development is conceptualized as the “fourth narcissistic blow” that science inflicts on humanity. We argue that this narcissistic blow has a potentially dramatic influence on perceptions of human society, interrelationships, and the self. We should, accordingly, expect dramatic changes in perceptions of the therapeutic act following the emergence of what we term the artificial third in the field of psychotherapy. The introduction of an artificial third marks a critical juncture, prompting us to ask the following important core questions that address two basic elements of critical thinking, namely, transparency and autonomy: (1) What is this new artificial presence in therapy relationships? (2) How does it reshape our perception of ourselves and our interpersonal dynamics? and (3) What remains of the irreplaceable human elements at the core of therapy? Given the ethical implications that arise from these questions, this paper proposes that the artificial third can be a valuable asset when applied with insight and ethical consideration, enhancing but not replacing the human touch in therapy.

Here are some thoughts:

This article conceptualizes the rise of generative artificial intelligence as the fourth narcissistic blow to human identity by challenging our unique monopoly over language. The authors introduce the concept of the artificial third to describe how technology enters the therapeutic space, transforming traditional interpersonal relationships. For psychologists, this paper is important because it shifts the conversation from technical efficiency to fundamental existential and ethical questions about autonomy, transparency, and the irreplaceable nature of human empathy. Insightfully, the study highlights that while artificial intelligence can process text, it lacks a true mind or subjective lived experience. Psychologists must therefore understand that this technology cannot replace the profound, nonverbal, emotional resonance of human connection. Ultimately, the article serves as a critical warning that embracing technology without maintaining strict ethical boundaries risks depersonalizing the therapeutic bond and undermining the very foundation of psychological healing.

A foundation model of vision, audition, and language for in-silico neuroscience

d’Ascoli, S., Rapin, J. et al. (2026)

Meta

Abstract

Cognitive neuroscience is fragmented into specialized models, each tailored to specific experimental paradigms, hence preventing a unified model of cognition in the human brain. Here, we introduce TRIBE v2, a tri-modal (video, audio and language) foundation model capable of predicting human brain activity in a variety of naturalistic and experimental conditions. Leveraging a unified dataset of over 1,000 hours of fMRI across 720 subjects, we demonstrate that our model accurately predicts high-resolution brain responses for novel stimuli, tasks and subjects, superseding traditional linear encoding models, delivering several-fold improvements in accuracy. Critically, TRIBE v2 enables in silico experimentation: tested on seminal visual and neuro-linguistic paradigms, it recovers a variety of results established by decades of empirical research. Finally, by extracting interpretable latent features, TRIBE v2 reveals the fine-grained topography of multisensory integration. These results establish artificial intelligence as a unifying framework for exploring the functional organization of the human brain.

A pdf can be found here.

Here are some thoughts:

This paper matters to psychologists because it introduces a single AI model capable of predicting brain responses across vision, language, and auditory processing simultaneously. Rather than relying on separate, task-specific models, TRIBE v2 offers a unified framework for understanding how the brain integrates multisensory information. It can replicate classic experimental findings without running new studies, potentially reducing the cost and time of psychological research. Its ability to generalize across hundreds of subjects also opens new possibilities for studying individual differences in cognitive and neural functioning.

Bio-Quantum Hybrid Linear Regression: A Novel Approach Combining Organoids Intelligence and Quantum Computing

Triana, H. (2026).

Research Gate

The paper is here.

Abstract

This paper introduces a novel Bio-Quantum Hybrid Linear Regression framework that integrates Organoids Intelligence (OI) with quantum computing operations to create a unified machine learning model. The proposed architecture combines two complementary computational paradigms: biological neural dynamics simulated through Brian2 [1], which models membrane potential evolution using differential equations, and quantum superposition operations implemented via Qiskit, which encode input values into qubit states through Y-rotation gates. The hybrid model performs linear regression by linearly combining outputs from both OI and quantum computing operations using learnable weights and a bias term, as formulated in yi = wqc·fqc(xi) + wOI·fOI (xi) + b. Experimental evaluation on synthetic datasets demonstrates the feasibility of integrating biological simulation  and quantum computing for regression tasks, while revealing important insights into the model’s behavior, limitations, and optimization requirements. The loss trajectory analysis shows increasing prediction errors without gradient-based optimization, highlighting the need for adaptive learning mechanisms. Despite current limitations, this work establishes a foundational framework for hybrid intelligence systems that leverage the complementary strengths of biological adaptive computation and quantum parallel processing capabilities. The paper also comprehensively discusses hardware and algorithmic limitations in both quantum computing (decoherence, qubit scalability, error correction) and organoid intelligence (scalability constraints, biological variability, ethical considerations), providing a roadmap for future research directions in hybrid computational intelligence that may transcend the constraints of traditional machine learning methodologies.

A pdf is here.

Here are some thoughts:

In essence, this paper represents a highly speculative, "blue-sky" proof of concept trying to answer a fundamental question: Can we plug a simulated biological brain and a quantum computer into the same mathematical equation?

While traditional AI relies entirely on silicon-based classical computing, the author is looking ahead to a distant future where we might outgrow standard microchips. By demonstrating that outputs from a simulated biological neuron and a simulated quantum qubit can be combined into a single formula, the paper attempts to lay a conceptual baseline for hybrid intelligence: systems that could theoretically pair the rapid, parallel problem-solving of quantum mechanics with the hyper-efficient, self-organizing adaptability of organic biology.  

However, the practical reality of the paper is a stark reminder of how far away that future is. Because the model lacked a basic learning mechanism to correct its mistakes, and because combining two highly unstable, noisy mediums (quantum states and biological cells) creates immense chaotic interference, the model completely failed to solve a basic math problem. Ultimately, the paper means that while bridging these two futuristic computational substrates is mathematically imaginable on paper, actually getting them to work together constructively is blocked by massive, unresolved engineering, algorithmic, and ethical barriers on both sides.

Educational Strategies for Clinical Supervision of Artificial Intelligence Use

Abdulnour, R. E., Gin, B., & Boscardin, C. K. (2025). 

New England Journal of Medicine, 393(8), 786–797.

https://doi.org/10.1056/nejmra2503232

Abstract

Many learners are more facile with the use of large language models in medicine than their supervisors are. The authors provide an approach to clinical supervision that can mitigate the perils and amplify the promise of AI.

The article is paywalled.

Here is how it opens:

Human–computer interactions have been occurring for decades, but recent technological developments in medical artificial intelligence (AI) have resulted in more effective and potentially more dangerous

interactions. Although the hype around AI resonates with previous technological revolutions, such as the development of the Internet and the electronic health record, the appearance of large language models (LLMs) seems different. LLMs can simulate knowledge generation and clinical reasoning with humanlike fluency, which gives them the appearance of agency and independent information processing. Therefore, AI has the capacity to fundamentally alter medical learning and practice. As in other professions, the use of AI in medical training could result in professionals who are highly efficient yet less capable of independent problem solving and critical evaluation than their pre-AI counterparts.

Here is a rather detailed summary:

This article provides a practical framework for supervising trainees who are using Artificial Intelligence (AI), specifically focusing on the risks to developing clinical reasoning skills. While the examples are medical, the core concepts of cognitive offloading, deskilling, and critical thinking are directly applicable to clinical psychology and psychotherapy supervision.

The Core Challenge: Balancing Efficiency with Skill Development

The authors argue that AI tools, particularly Large Language Models (LLMs), present a paradox. They can enhance learning through simulation and cognitive offloading of rote tasks, but they also pose significant risks when used to replace, rather than augment, complex clinical reasoning. The central concern is that over-reliance on AI for tasks like diagnosis, case formulation, or treatment planning can lead to:

• Deskilling: Loss of newly acquired clinical reasoning skills.

• Never-skilling: Failure to develop essential competencies in the first place.

• Mis-skilling: Reinforcement of incorrect or biased clinical behavior due to flawed AI output.

This is especially dangerous because AI operates as a "black box," generating persuasive but potentially biased or inaccurate responses without transparent reasoning.

The "Leap of Faith" and the Supervisor's Role

A key concept is the AI interaction: a moment when a clinician receives an AI-generated judgment that cannot be fully retraced, requiring a "leap of faith" to trust it. The supervisor's job is to teach trainees to recognize these moments and pause for critical evaluation, rather than passively accepting the output.

The supervisor-learner dynamic may be inverted, as trainees are often more adept with the technology. The article reframes this as a shared learning opportunity, where supervisors and learners co-explore AI's capabilities and limitations in a "community of practice."

The DEFT-AI Framework for Supervision

The authors propose a structured, stepwise approach called DEFT-AI (Diagnosis, Evidence, Feedback, Teaching, and AI Recommendation) to turn an AI interaction into an educational moment that builds critical thinking. Here is how it can be applied in a psychology context:

• Diagnosis, Discussion, and Discourse: The supervisor asks the trainee to verbalize their own clinical reasoning before revealing the AI's input. Questions include: "What is your formulation and differential? What prompts did you use with the AI? Did the AI's output change your thinking, and how?"

• Evidence: The supervisor probes the trainee’s ability to support their clinical reasoning with psychological theory, evidence-based practice, and knowledge of the patient’s unique context. Simultaneously, the supervisor probes AI literacy: "How do you think the AI reached this conclusion? What are the known biases or weaknesses of this tool for this specific clinical question?"

• Feedback: The supervisor guides the trainee in self-reflection on gaps in their clinical knowledge, potential biases, and their interaction with the AI tool.

• Teaching: The supervisor provides targeted teaching to address identified gaps, reinforcing foundational clinical reasoning and AI literacy.

• AI Engagement Recommendation: The supervisor makes a clear recommendation on the appropriate future use of AI for the trainee, ranging from supervised practice to independent use with self-monitoring.

Cyborg vs. Centaur: Two Styles of AI Use

The article identifies two distinct collaboration styles that supervisors should help trainees recognize and shift between:

• Centaur Strategy: A strategic division of labor. The human delegates specific, well-defined tasks to AI (e.g., drafting psychoeducational materials, summarizing session notes) but relies on their own clinical judgment for core tasks like diagnosis and treatment planning. This is the preferred strategy for high-risk tasks.

• Cyborg Strategy: A tight, iterative interweaving with AI for every step of a task (e.g., co-constructing a case formulation by prompting, correcting, and refining with an LLM). This is efficient for low-risk, creative, or well-defined tasks but carries a high risk of deskilling.

Adaptive AI practice is the ability to fluidly switch between centaur, cyborg, and AI-independent modes based on the complexity and risk of the clinical task at hand.

Promoting AI Literacy: The "Verify and Trust" Paradigm

Ultimately, the goal is to foster a "verify and trust" mindset over blind trust. Supervisors must teach two key skills:

1. Critical Appraisal of AI Output: Trainees must independently acquire and appraise evidence (e.g., clinical guidelines, therapeutic literature) for a clinical question and compare their own conclusions to the AI's output before accepting it.
2. Effective Prompting: Trainees need to learn how to craft specific, context-rich, and unbiased prompts. Techniques like asking the AI to "think out loud" (chain-of-thought prompting) can expose the AI's reasoning and facilitate critical assessment.

For psychologists and clinical supervisors, this framework offers a clear, theory-grounded method to proactively integrate AI into supervision while safeguarding the development of independent, adaptive, and critical clinical judgment in trainees.

Living intelligence toward human-level models (HLMs) via Organoid-AI integration

Bai, L., Wang, J., Lai, Y., & Su, J. (2025).

EngMedicine, 2(4), 100106.

https://doi.org/10.1016/j.engmed.2025.100106

Abstract

The convergence of brain organoids and artificial intelligence (AI) has driven the development of organoid intelligence (OI), a new paradigm for constructing human-level cognitive models. Brain organoids derived from human stem cells exhibit self-organizing neural networks with dynamic activity and plasticity, offering a biologically based alternative to conventional AI systems. The integration of living networks with computational frameworks enables the design of closed-loop systems that combine the adaptability of biological tissues with the scalability and interpretability of AI. This approach not only provides a novel model for studying human cognition but also opens new pathways for biologically inspired computing. The development of such hybrid systems requires interdisciplinary collaboration among stem cell biology, bioengineering, neuroscience, and machine learning. The long-term goal is to establish biohybrid platforms capable of learning, memory formation, and task-specific computation, thereby redefining our understanding of intelligence and enabling the next generation of neurotechnologies.

Highlights

• Organoid Intelligence (OI) combines brain organoids and AI.

• OI creates biologically embodied models for human-level cognition.

• Biohybrid platforms can learn, remember, and perform computations.

• OI requires interdisciplinary collaboration for development.

Here are some general thoughts:

We are witnessing the infancy of true synthetic biological intelligence. While current applications are constrained to pattern recognition and disease modeling, the long-term trajectory completely disrupts the binary view of technology as "artificial" and biology as "natural." It forces tech developers and ethicists alike to confront a reality where the next generation of advanced intelligence might not be coded, but grown.

New 3D device harnesses living brain cells for computing

Princeton University

Office of Engineering

Originally posted April 27, 2026

Princeton researchers have combined brain cells and advanced electronics into a 3D device that can be programmed to recognize patterns using computational techniques.

Past attempts at using brain cells to do computation have relied on 2D cultures grown in a petri dish or 3D clusters that are probed and monitored from outside. The Princeton device takes a different approach, working from the inside out.

Using advanced fabrication techniques, the team created a 3D mesh made of microscopic metal wires and electrodes supported by a thin epoxy coating. Because the coating is so thin, it has just the right amount of flexibility to interface with the soft neurons that grow around it. The team used the mesh as a scaffold to culture tens of thousands of neurons into a vast 3D network that can be used to do computation.

The info is here.

The research paper is here.

Here are some thoughts:

Princeton University researchers have developed an innovative 3D device that integrates roughly 70,000 living biological neurons with advanced electronics to perform computational tasks, such as recognizing spatial and temporal electrical pulse patterns. Published in Nature Electronics, the study details a novel "inside-out" approach where an ultra-thin, flexible epoxy-coated mesh of microscopic metal wires and electrodes serves as a scaffold for the soft brain cells to grow around, allowing scientists to record and stimulate electrical activity at an unprecedentedly fine scale. By tracking and manipulating these neural connections over a six-month period, the team successfully trained an algorithm to distinguish between different pattern inputs, demonstrating a crucial first step toward creating highly energy-efficient 3D biological neural networks that could eventually alleviate the immense power demands of modern AI while providing deeper insights into neuroscience and neurological diseases.

Benchmarking Large Language Models Against Psychiatry Residents Using Traditional Institutional Assessments

Sethi, M. I. S. et al. (2026).

Indian Journal of Psychological Medicine, 

02537176261435658.

https://doi.org/10.1177/02537176261435658

Background:Artificial intelligence (AI) models demonstrate remarkable capabilities in healthcare applications, yet their performance compared to medical trainees in psychiatric education remains unexplored. This study evaluated the comparative performance of large language models (LLMs) against first-year psychiatry residents in standardized assessments at a premier Indian medical educational institute.

Methods:For this study, the already-scored answer sheets for Theory Papers I and II, as well as unmanned, non-interactive Objective Structured Clinical Examinations (OSCEs) with image-based tasks, from all 25 first-year psychiatry residents (March 2024 exam) were obtained from the examination section of the institute. The same question papers were then uploaded into three AI models (ChatGPT−3.5, Gemini Advanced, and Claude Sonnet). Four blinded faculty members evaluated the responses generated by the AI models. Final, the scores of the AI models and psychiatry residents were analyzed for comparison. Statistical analysis employed Kruskal–Wallis tests with post hoc Mann–Whitney U comparisons.

Results:AI models outperformed residents in theoretical assessments. In Paper I (theory), AI models achieved mean scores (standard deviation) of Claude Sonnet 67.88 (10.63), ChatGPT−3.5 70.38 (3.95), and Gemini Advanced 71.25 (3.86), compared to residents’ 58.0 (2.58). Paper II (theory) assessments showed even larger gaps, with AI models scoring Claude Sonnet 72.88 (3.77), ChatGPT−3.5 71.0 (3.56), and Gemini Advanced 69.63 (12.86), compared to residents’ 50.96 (2.49). OSCE performance patterns differed markedly. Paper I OSCEs showed equivalent performance: AI: 13.0; residents’: 13.16 (1.49), while Paper II OSCEs revealed variable results: Claude Sonnet excelled at 20.0 (1.41), but ChatGPT−3.5 underperformed at 15.0 (0.50), compared to residents at 16.6 (1.55). Inter-rater reliability coefficients remained excellent ( intraclass correlation coefficients [ICC]: 0.810–0.934).

Conclusions:While AI demonstrated superior theoretical knowledge, equivalent or variable practical skills performance reveals fundamental limitations in clinical reasoning and contextual understanding. These findings necessitate reconceptualizing psychiatric education to emphasize uniquely human competencies while leveraging AI’s capabilities for knowledge synthesis.

Here are some thoughts:

This study compared three large language models (LLMs) to first-year psychiatry residents using real institutional exams in India. The LLMs consistently outperformed residents on theoretical assessments (by 17–43%) but showed equivalent or inconsistent performance on practical OSCEs, revealing critical gaps in clinical reasoning and cultural contextualization. The authors conclude that psychiatric education should shift focus toward uniquely human skills like empathy and judgment, while using AI as a tool for knowledge synthesis.