Journal of Mental Health Research and Innovation

INTRODUCTION

Background

The advent of antiretroviral therapy has increased hope for life preservation among people living with HIV (PLWH); however, the diagnosis of HIV continues to have substantial psychological and social consequences. Evidence indicates that quality of life may be significantly affected from the point of diagnosis, reflecting the psychological and social impact of living with the condition [1]. The period immediately following diagnosis is therefore a particularly important window for the provision of appropriate psychosocial support.

Counselling provided by healthcare providers with little or no prior experience in the field of mental health is often relied upon as a form of psychological first aid for newly diagnosed PLWH [2]. However, conventional counselling may be insufficient to mitigate the exceptional psychological strain experienced during this critical period. In low- and middle-income countries (LMICs) such as Nigeria, this challenge is further compounded by the continuous emigration of healthcare professionals and an inadequate doctor-to-patient ratio, making interventions that are heavily dependent on human resources difficult to scale and deliver in a timely manner [3].

Artificial intelligence (AI) is increasingly being applied within mental healthcare and has emerged as a potential tool for addressing some of these service-delivery challenges. Its applications include predictive analysis, digital counselling, emotion recognition, and patient engagement [4]. AI-based psychosocial interventions, including chatbots, algorithm-based screening systems, and virtual mental health assistants, have demonstrated potential for mitigating anxiety and stress at the global level [5]. Such technologies may provide accessible and scalable avenues for psychosocial support, particularly in settings where mental health resources are limited [6].

Despite these developments, there remains a paucity of synthesized evidence regarding the integration of AI into psychosocial support systems for PLWH, particularly in relation to anxiety, stress, and psychological distress. This gap is especially relevant to LMICs such as Nigeria, where the contextual applicability of technologies developed in other settings remains uncertain [7]. Differences in cultural and linguistic contexts, digital literacy, technological infrastructure, accessibility, and healthcare systems may influence the acceptability, feasibility, and effectiveness of AI-based psychosocial interventions [8].

A scoping review can therefore provide a comprehensive mapping of the existing evidence, including the range of AI-based psychosocial interventions that have been studied, their target populations, reported outcomes, strengths and limitations, and gaps in the current literature. Such evidence is particularly important for determining how existing AI-based interventions may be adapted to the Nigerian context and potentially integrated into HIV and mental health care systems [9].

This study therefore sought to map the available evidence on the integration of artificial intelligence into psychosocial support systems for people living with HIV, with particular attention to interventions addressing anxiety and stress. It further aimed to identify contextual and methodological gaps in the literature and explore the potential implications of existing evidence for the development of locally relevant, culturally appropriate, and context-specific AI-based psychosocial support systems within HIV care in Nigeria.

Research Questions

• What evidence exists on AI-based psychosocial interventions addressing anxiety disorder and
  stress in people living with HIV?
• What gaps exist in current literature as regards adapting these available AI systems in the
  context of Nigerian mental health and HIV care?
• How have AI tools (e.g., chat boxes, virtual counseling, predictive algorithms) been used to
  address psychosocial distress in the general population of people living with HIV?
• What are the strengths, limitations, and reported outcomes of these AI-based psychosocial
  support systems?
• How can future research improve the quality and relevance of evidence on artificial
  intelligence integration in psychosocial support systems of vulnerable populations?

Aim of the Study

The aim of this study is to provide locally relevant, context-specific data that will encourage the
integration of artificial intelligence into psychosocial support systems, preventing anxiety
disorder and stress in newly diagnosed people living with HIV (PLWH).

Objectives

To identify the contextual and methodological gaps in existing literature relevant to
integrating artificial intelligence-based psychosocial intervention to prevent anxiety disorder
and stress in people living with HIV

• To investigate how AI tools have been used to address psychosocial distress in the general population of people living with HIV.
• To identify the strengths, limitations, and reported outcomes of these AI-based psychosocial support systems.
• To propose how future research can improve the quality and relevance of evidence on artificial intelligence integration into psychosocial support systems of vulnerable populations.

MATERIALS AND METHODS

A scoping review was conducted following the PRISMA-SCR guidelines and using the Arksey and O’Malley methodological framework.

Electronic databases including PubMed, Google Scholar, HINARI, SCOPUS, and PsycINFO

Reference list of relevant papers.

Search strategy:

Searches were conducted using strings of keywords formed with Boolean operators.

Combinations include:

("Artificial Intelligence" OR "machine learning" OR "digital mental health") AND ("psychosocial support" OR "mental health intervention") AND ("stress" OR "anxiety") AND ("HIV" OR "PLHIV").

Studies exploring AI-based psychosocial interventions (qualitative, quantitative, or mixed- methods).

Studies addressing mental health, anxiety, or stress among PLHIV with the aid of digital tools/technology.

Studies within 10 years of publication written in English (2015-2025).

A data-charting form (literature review matrix) capturing details of the author, country,AI-based intervention used, target population, outcomes, and limitations was used.

The data was then analyzed thematically to identify major trends, target areas for intervention, and gaps in literature

A scientific mapping of current evidence on AI-integrated psychosocial support systems for mental health disorders (especially anxiety and stress) in HIV populations. Identification of contextual and methodological literature gaps in LMICs such as Nigeria. A theoretical framework revealing potential ways for integrating AI-linked psychosocial support systems into HIV care programs of the newly diagnosed and a scholarly foundation for future empirical studies and policy formulation.

Day 1: Set up and development of database search strategy Literature search protocol

Day 2: Initial screening of research topics and abstracts Studies shortlisted

Day 3: Full-text review of selected topics and data extraction Literature review matrix

Day 4: Drafting of key findings Thematic synthesis of Analysis and summary

Day 5: Compilation of final abstract and proposal submission Completed research document.

Several literatures were reviewed to explore the integration of artificial intelligence in psychosocial support systems of people living with HIV (PLWH), searching databases such as Google Scholar, HINARI, SCOPUS, PubMed, and Psych INFO. The database were searched for articles related to the study's theme and specific articles selected are those that were relevant to our research objectives. Further searches were conducted using the following keywords:"artificialintelligence","psychosocialsupport","anxiety","stress","HIV","scopingreview.

Articles published between 2015 and 2025 were manually sorted ensuring no bias in the selection process. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) checklist was used in this study.

Inclusion criteria : Studies centered on the main topic and utilizing relevant terminology.

Exclusion criteria : Research on similarity indices that didn't align with our topic of interest.

Studies on AI-tools that required human support.

PRISMA-ScR Flow Chart

LITERATURE REVIEW

RESULTS AND DISCUSSION

Geographic Diversity of Included Studies

This bar chart visualizes the distribution of the 17 reviewed studies by region, based on the

Estimated Predominance of AI Tools

This pie chart illustrates the relative focus on different types of AI tools identified in the studies, showing that chatbots and mobile applications are the predominant modalities. The percentages are estimated based on the textual emphasis.

Severity of Key Limitations

This bar chart provides a qualitative assessment of the methodological and contextual limitations discussed in the review, based on the emphasis given in the text. The highest-rated limitations concern the study design (feasibility/pilot) and the lack of localization for LMICs.

Impact of Key Strengths

This bar chart summarizes the positive attributes of the AI interventions, indicating that improved accessibility, engagement, and user acceptability are the highest-impact strengths identified across the reviewed studies.

Overview of Included Studies

In total, 17 studies were reviewed to explore the evidence on artificial intelligence (AI)-based psychosocial interventions addressing anxiety and stress among people living with HIV (PLWH). The studies spanned across continents including Asia (China, Malaysia, Hong Kong), Africa (South Africa, Ghana), North America (Canada, USA), and Europe (Italy, Spain, UK). There was however no study originating from Nigeria [8]. Majority of the studies were conducted between 2019 and 2025, showing the rise with time of contemporary evidence on the relevance of AI in mental health care [10].

The AI tools identified included chatbots, mobile health (mHealth) applications, machine learning programs, and digital conversational agents designed to promote adherence, self-testing, or psychological wellbeing [11]. People living with HIV (PLWH), men who have sex with men (MSM), and young people at risk of HIV infection made up the larger proportion of the target population in the studies. 

Evidence on AI-Based Psychosocial Interventions for Anxiety and Stress in PLWH

The potential of AI-driven interventions to address psychosocial distress was demonstrated by multiple studies [12]. For instance, Wang evaluated the AICare hybrid intervention in China and reported improvements in disease adaptation, quality of life, and reduction of anxiety symptoms among [1]. Similarly, Barroso, found that a Cognitive Behavioral Stress Management (CBSM) mobile application enhanced coping skills and was well accepted by users [13].

Digital interventions utilizing mental health applications, such as Wyssa and CBT-based mHealth tools, were shown to effectively reduce symptoms of depression and anxiety in low- and middle-income countries (LMICs) [14]. Moreover, the systematic review by Adepoju likewise Adjorlolo in Sub-Saharan Africa highlighted promising evidence that AI-based psychosocial therapies and digital psychological interventions, can improve mental health outcomes and adherence to antiretroviral therapy (ART) [15]. 

It was however observed that most of the interventions were feasibility or pilot studies, with limited randomized controlled trials (RCTs) measuring anxiety and stress as primary outcomes. Clinical efficacy of interventions was not well emphasized with many studies rather focusing on usability, feasibility, and user engagement [15]. Thus, while AI-based psychosocial interventions demonstrate strong potential, there is paucity of robust empirical evidence.

AI Tools Applied in Psychosocial Support for PLWH

Chatbots were the predominant AI-tool identified in this study. The NoIwazi and MARVIN chatbots in South Africa and Canada respectively were designed to deliver HIV self-testing support and psychosocial counseling. Worthy of note is the Nolwazi bot local language (IsiZulu) interface and conversational tone, which enhanced user acceptability and trust [16]. Participants reported feeling understood “as if talking to a real person,” highlighting the importance of cultural adaptation in building AI systems.

Similarly, users of MARVIN and REALbot had round the clock access to expert-validated information and counseling, encouraging engagement and self-management [17]. The predictive analytics potential of mental health surveillance by machine learning algorithms were also used to analyze behavioral and emotional expressions in online text data from HIV-related social networks [13].

Predictive AI systems and virtual counseling algorithms however remain underutilized when compared to chatbots in spite of these advancements [18]. This underutilization of diverse AI tools suggests a primitive era of integration of artificial intelligence into psychosocial support frameworks for PLWH [19].

Strengths and Limitations of AI-Based Psychosocial Systems

Strengths demonstrated in the review include:

Improved accessibility and engagement: Especially suitable for stigmatized population as AI tools provide all day long support and confidentiality [20].

Feasibility and user acceptability : Positive user experiences and high feasibility scores were consistently demonstrated across many studies [2].

Cultural adaptability: AI systems that integrated local languages and contextual cues, (e.g., NoIwazi chatbot) significantly improved acceptability [16].

Integration of evidence-based frameworks: Better psychosocial outcomes were observed with tools that incorporated evidence-based principles like cognitive behavioral therapy (CBT) and dialectical behavioural therapy (DBT) [19].

Limitations that were evident include:

Methodological constraints: Only a few studies employed control groups or long-term follow ups,most of the others were cross-sectional or pilot designs [21].

Limited outcome measurement: Studies measuring anxiety and stress as primary outcomes were rare, HIV self-testing, engagement, or adherence is often the focus.

Technological and infrastructural barriers: As observed in many of the studies, the interventions required high internet connectivity, smartphone ownership, and digital literacy [6].

Lack of localization for LMICs: High and middle-income countries dominate in terms of AI-tools development, reducing their adaptation and contextual relevance to African settings [22].

Ethical and data privacy concerns: As shown in several studies, AI-mediated counseling holds potential confidentiality risks and mistrust [4].

These limitations emphasize that despite the promising potentials of AI-systems, their evidence base remains primitive, and their adaptation in the context of LMIC like Nigeria is urgently needed [8].

Gaps in Literature Relevant to Nigeria

The most critical gap is the paucity of contextual evidence evaluating AI-based psychosocial interventions for PLWH in Nigeria [6]. Despite a high burden of HIV and increasing digital health interest there was no identifiable study conducted in Nigeria during the course of this review. This emphasizes the need for locally driven research that integrates cultural, linguistic, and infrastructural factors [8].

It was also observed that there is lack of integration of AI tools into existing psychosocial support systems within HIV clinics. Most reviewed interventions were digital platforms operating independently with insufficient link to mental health professionals or ART adherence counselors [23]. Also, it was difficult to generalize findings in the context of mental health as only few interventions assessed stress and anxiety outcomes quantitatively. Lastly, there is a poor articulation of data policies and ethical frameworks for AI in health across African studies, this poses a barrier to the acceptability and sustainability of digital interventions [6].

Implications for Future Research

There is need for the development of AI-based psychosocial interventions that are culturally and linguistically adapted to Nigeria, Future studies should focus on integrating them into routine HIV care and mental health services [8]. Rigorous study designs, such as randomized controlled trials (RCTs), must be prioritized in doing this while employing validated measures of assessing psychological outcomes such as stress and anxiety among PLWH. Inter-professional collaborations between psychiatrists, computer scientists, behavioral researchers, and policy makers are also necessary to design scalable and ethical AI systems [20]. Furthermore, digital health literacy and capacity building among both clinicians and clients will enhance acceptability and sustainability [8]. Lastly, predictive algorithms and sentiment analysis tools should be incorporated in HIV care in order to enable early identification of psychosocial distress and develop preventive mental health strategies [24].

SUMMARY AND CONCLUSION

Summary of Findings

This review systematically analysed current evidence on Artificial Intelligence (AI)-based psychosocial interventions aimed at addressing anxiety and stress among people living with HIV (PLWH). It also reveals contextual and methodological gaps relevant to the integration of these interventions in the Nigeria's mental health framework and HIV care clinics. 17 Literature published between 2015 and 2025 were thoroughly reviewed, they covered regions including China, South Africa, Canada, the United States, Malaysia, Ghana, Spain, Italy and the United Kingdom. No study originated from Nigeria despite the progress of digital health globally, highlighting a significant regional research gap.

The studies revealed that AI-based interventions notably chatbots, mobile health mHealth applications, and machine learning tools were increasingly employed to provide mental health and psychosocial support to PLWH and populations at risk of HIV infection. These tools were beneficial in terms of facilitating HIV self-testing, providing digital counseling, treatment monitoring, and emotional well-being support.

Majority of the interventions demonstrated positive psychosocial impact reduced anxiety, improved coping mechanisms, and better engagement with HIV care were noted. For example, the AI Care hybrid intervention improved disease adaptation and quality of life among PLWH, while the Cognitive Behavioral Stress Management (CBSM) mHealth App enhanced stress coping mechanisms. In addition, DBT-informed mHealth tools were effective regulating emotion and reducing depressive symptoms among men living with HIV in China. The evidence base remains limited despite these promising benefits due to methodological inadequacies, including small sample sizes, feasibility-focused designs, and lack of rigorous randomized controlled trials (RCTs). Also, while most studies focused on usability and feasibility, clinically validated mental health outcomes such as stress and anxiety levels were not well prioritized.

From the NoIwazi chatbot report in South Africa, it was shown that high user satisfaction was achieved because it communicated in the local IsiZulu language and imitated participants’ conversational tone. Hence, cultural and linguistic contextualization emerged as a critical success factor. This underscores the importance of designing AI systems that align with users’ cultural identity and communication patterns a key point for potential adaptation in Nigeria.

Absence of studies conducted in Nigeria reveals significant gaps in both clinical practice and research. While the country faces a high burden of HIV and rising cases of psychosocial distress among newly diagnosed individuals, there is little empirical data to support AI integration into psychosocial support systems [8].

CONCLUSION

According to findings presented in this study, AI-based psychosocial interventions have a lot of potential in offering support to PLWH regarding their mental and emotional health. Interventions such as AI chatbots, virtual counseling platforms, and mHealth applications have demonstrated feasibility, potential efficacy and acceptance in enhancing coping skills, decreasing psychological distress, and improving ART adherence. However, there were no locally specific, non-preliminary evidence supporting these benefits. As most of the studies originate from high-income countries, limited representation from low-income countries like Nigeria was also conspicuous. This disparity suggests the urgent need for contextually adapted AI-driven interventions that incorporates linguistic diversity, digital literacy, and infrastructural realities.

It also shows how mental health service gap can be bridged by AI integration into psychosocial support systems for HIV care in Nigeria by providing accessible, confidential, and scalable interventions. With Nigeria’s expanding digital health policies and smartphone utility AI-driven solutions could complement human counseling and betterment of care.

Nevertheless, inter- sectoral collaboration between mental health professionals, computer scientists, public health experts, and policymaker will contribute to successful implementation. Also, ethical considerations including data privacy, informed consent, algorithmic fairness, and accountability must be addressed to ensure confidentiality and boost responsible AI deployment.

To conclude, AI-based psychosocial interventions offer a game-changing chance to transform the HIV care and mental health landscape in Nigeria. Considering how healthcare workers in Nigeria and other LMIC can benefit from the reduced burnout and improved mental health status among people living with HIV (PLWH), adapting global innovations to local realities will enhance the resilience of the healthcare system. 

RECOMMENDATIONS

Research Recommendations

Conduct Context-Specific Studies in Nigeria:

Future studies about AI-based psychosocial intervention should be tailored to Nigerian sociocultural and linguistic contexts. Tools should be sensitive to the local communication styles and should incorporate native languages (e.g. Hausa,Igbo,Yoruba).

Adopt Rigorous Research Designs: National health policy and evidence base will be strengthened when empirical studies move beyond feasibility to randomised controlled trials (RCTs) with validated assessments of psychological outcomes (stress, anxiety and coping).

Integrate AI Tools within Clinical and Community-Based Psychosocial Systems: Rather than being isolated digital platforms, the sustainability of care can be promoted when AI tools are embedded into existing psychosocial support systems and ART adherence programs within HIV care clinics.

Advance Predictive and Personalized AI Models: Proactive mental health interventions can be promoted when future research go beyond chatbots and explore machine learning algorithm that can detect early cues of psychosocial distress from text, voice or behavioral data.

Ethical and Data Governance Frameworks: National ethical guidelines should be established for AI use in mental health and HIV care, with focus on confidentiality, and data protection to maintain patient security and compliance [23].

Policy and Practice Recommendations

Incorporate AI into Nigeria’s Digital Health Strategy: AI-driven mental health interventions should be prioritized by the Federal Ministry of Health and National Agency for the Control of AIDS (NACA) to encourage broader digital health transformation.

Capacity Building for Health Professionals: Health practitioners and HIV counselors should be equiped with basic AI literacy, by developing training programs in support of this, they can be enabled to collaborate in the design and deployment of artificial intelligence.

Public-Private and Academic Partnerships: Universities, health agencies, and technology firms need to collaborate to create locally adaptable and affordable AI solutions for psychosocial support in HIV care.

Enhance Accessibility through Mobile Integration: To reach rural and underserved populations, AI-based psychosocial support tools should be optimized for offline functionality on smartphones by using mobile platforms requiring minimum or no data subscription.

Monitor and Evaluate Implementation Outcomes : Standard monitoring and evaluation frameworks should be established to assess effectiveness, compliance to ethics, and user satisfaction of AI-based interventions in a real Nigerian settings.

Implications for Psychiatry and Mental Health

AI can serve as a complementary tool for mental health professionals; it should not be seen as a total replacement. This way, service delivery can be improved and adherence to treatment can be improved among PLWH. The integration of AI into psychosocial support systems presents an evolving model in the field of psychiatry, thus opening an avenue for digital psychiatry which promotes remote psychological assessment and intervention in resource limited environments.

Final Remark

Artificial intelligence offers a unique opportunity to bridge the gap between mental health needs and service delivery among PLWH in Nigeria, including offering the much required psychological first aid at the time of diagnosis. Building on global evidence, future locally driven, ethically guided, and culturally attuned AI interventions can strengthen psychosocial resilience, prevent anxiety and stress, and ultimately improve the quality of life of people living with HIV.

ETHICAL CONSIDERATIONS

This study is a retrospective analysis of published datasets, hence ethical approval is not required. We however ensured proper citation, transparency and adherence to PRISMA-SCR reporting standards.