Summary: New research shows that an AI assistant can conduct psychological screening interviews with greater diagnostic accuracy than widely used mental health scales. In a sample of 303 participants with confirmed mental conditions, the AI Assistant Alba provided diagnostic suggestions made for the DSM after a short interview, rating scales for eight out of eight disorders.
Alba was particularly successful at differentiating conditions that often overlap, such as depression and anxiety, that traditional scales tend to score similarly. Participants also report a positive user experience, demonstrating that adaptive AI can provide a crowdsourced, human-centered tool to support a critical health pathway while maintaining the important role of the Clinician.
Key facts:
- High Accuracy: AI-based interviews based on standard rating scales in eight psychological tests.
- The best difference: Alba is clearly divided between betting symptoms such as anxiety and depression.
- Good Experience: Participants reported that the AI interview felt empathetic, supportive and engaging.
Source: Lund University
New research shows that an AI assistant can conduct diagnostic interviews with patients with higher accuracy than the scales used in healthcare today. In the study, 303 participants were interviewed about AI Alba's assistant, who suggested a psychiatric diagnosis.
In addition to being interviewed by the AI assistant, participants also had to fill out standard rating scales for standard psychiatric diagnoses.
The results showed that the AI assistant assessment was more consistent with the actual diagnosis of the participants than the rating scales.
The study included people with a confirmed diagnosis of conditions such as depression, anxiety, obsessive-compulsive disorder, ADHD, autism disorder, and a group of depression.
All participants had an online interview with the AI assistant, Alba, asking him 15-20 open-ended questions about their mental health and a proposed diagnosis based on the DSM-5 – the International Manual of Psychiatric Diagnosis.
The AI assistant achieved high diagnostic accuracy in eight out of nine diagnoses, and could clearly distinguish between diagnoses that often overlap. For example, standard rating scales tend to give similar readings for depression and anxiety, whereas Alba's test can identify conditions more clearly.
Participants also describe the user experience as positive – many see the AI assistant as Muntuc, relevant and dynamic.
“The interview that can be done in a safe place at home before meeting the clinic is of great value. The results point to the perfection of psychology, without replacing the research at Lund University and the Founder of the company, talk to alba.
It analyzes the entire diagnostic manual – not just individual cases
According to Sverker Sikström, the study marks a clear step forward in the research of diagnostic tools for mental health. Previous studies have often been limited to linking individual diagnoses or lack of clarity based on diagnostic criteria, whereas Alba can propose and justify all diagnoses included in the DSM Manual.
TRUTH BOX: What does it mean to talk to alba?
Talk with Alba is an online AI tool for assessment, treatment and management of mental health for professionals (psychiatrists, psychizatrists and doctors) working in this area.
The tool includes AI-Powered Clinical Interviews and CBT for patients, automatic detection of appropriate mental health according to DSM-5, information with clinical patient interviews, and writing and journaling of patient meetings.
Alba, which is used in clinics in Sweden and other countries, is owned by Talktoalba AB.
Important Questions Answered:
A: Yes. The AI assistant showed higher diagnostic accuracy than the measured scale in eight of the nine cases.
A: It does. Alba differentiated diagnoses such as anxiety and depression more clearly than conventional instruments.
A: Most described the experience as compassionate, appropriate, and supportive.
Editing notes:
- This article was edited by the editor of neuroscience news.
- The journal is fully reviewed.
- Additional context added by our staff.
About this research on AI and Mental Health
Author: Lotte billing
Source: Lund University
Contact: Lotte Bill – Lund University
Image: This photo is posted in Neuroscience News
Actual research: Open access.
“AI-dedicated discussion of mental health and mental health” by Sverker Sikström et al. Scientific Reports
-Catshangwa
AI-assisted counseling in mental health
Standard mental health evaluations often include clinical interviews conducted by highly trained clinicians. While effective, this approach faces significant limitations, including high cost, high clinical burden, technical diversity, and lack of specificity.
Recent advances in large-scale linguistic models (LLMS) offer a promising avenue to address these limitations by simulating clinically controlled conversations with AI systems.
However, few studies have rigorously validated such tools. In this study, we used Talktoalba to develop and analyze an AI assistant designed to conduct clinical interviews consistent with DSM-5 criteria.
Participants (Ni= 303) It included people with a diagnosed mental health disorder, that is, major depressive disorder (MDD), obsessive-compulsive disorder (hyperactivity disorder (ASD), eating disorder (ED), eating disorder, substance abuse disorder, substance abuse disorder (Sud), and bipolar disorder (BD) – Made healthy healthy controls.
An AI assistant conducted diagnostic interviews and assessed the probability of each disorder, while another AI program analyzed the interview to confirm the diagnostic methods and generate comprehensive reasons for its conclusions.
The results revealed that the AI-Powered Clinical Interview found high agreement (ie, Cohen's Kappa), sensitivity, and specificity for self-reported, drug-disease disorders compared to measurement scales.
It also showed high interdependence between diagnostic categories. In addition, the majority of participants rated high-powered conversation as empathetic, appropriate, understanding, and supportive.
These findings suggest that AI-powered clinical interviews can be accurate, standardized, and powerful diagnostic tools for diagnosing common mental disorders.
Their flexibility, low cost, and good user experience position them as an important alternative to traditional diagnostic methods, with the potential to apply to social health care.
