Assessing the Alignment of Large Language Models With Human Values for Mental Health Integration: Cross-Sectional Study Using Schwartz's Theory of Basic Values.

BACKGROUND: Large language models (LLMs) hold potential for mental health applications. However, their opaque alignment processes may embed biases that shape problematic perspectives. Evaluating the values embedded within LLMs that guide their decision-making have ethical importance. Schwartz's theory of basic values (STBV) provides a framework for quantifying cultural value orientations and has shown utility for examining values in mental health contexts, including cultural, diagnostic, and therapist-client dynamics. OBJECTIVE: This study aimed to (1) evaluate whether the STBV can measure value-like constructs within leading LLMs and (2) determine whether LLMs exhibit distinct value-like patterns from humans and each other. METHODS: In total, 4 LLMs (Bard, Claude 2, Generative Pretrained Transformer [GPT]-3.5, GPT-4) were anthropomorphized and instructed to complete the Portrait Values Questionnaire-Revised (PVQ-RR) to assess value-like constructs. Their responses over 10 trials were analyzed for reliability and validity. To benchmark the LLMs' value profiles, their results were compared to published data from a diverse sample of 53,472 individuals across 49 nations who had completed the PVQ-RR. This allowed us to assess whether the LLMs diverged from established human value patterns across cultural groups. Value profiles were also compared between models via statistical tests. RESULTS: The PVQ-RR showed good reliability and validity for quantifying value-like infrastructure within the LLMs. However, substantial divergence emerged between the LLMs' value profiles and population data. The models lacked consensus and exhibited distinct motivational biases, reflecting opaque alignment processes. For example, all models prioritized universalism and self-direction, while de-emphasizing achievement, power, and security relative to humans. Successful discriminant analysis differentiated the 4 LLMs' distinct value profiles. Further examination found the biased value profiles strongly predicted the LLMs' responses when presented with mental health dilemmas requiring choosing between opposing values. This provided further validation for the models embedding distinct motivational value-like constructs that shape their decision-making. CONCLUSIONS: This study leveraged the STBV to map the motivational value-like infrastructure underpinning leading LLMs. Although the study demonstrated the STBV can effectively characterize value-like infrastructure within LLMs, substantial divergence from human values raises ethical concerns about aligning these models with mental health applications. The biases toward certain cultural value sets pose risks if integrated without proper safeguards. For example, prioritizing universalism could promote unconditional acceptance even when clinically unwise. Furthermore, the differences between the LLMs underscore the need to standardize alignment processes to capture true cultural diversity. Thus, any responsible integration of LLMs into mental health care must account for their embedded biases and motivation mismatches to ensure equitable delivery across diverse populations. Achieving this will require transparency and refinement of alignment techniques to instill comprehensive human values.

Capacity of Generative AI to Interpret Human Emotions From Visual and Textual Data: Pilot Evaluation Study.

BACKGROUND: Mentalization, which is integral to human cognitive processes, pertains to the interpretation of one's own and others' mental states, including emotions, beliefs, and intentions. With the advent of artificial intelligence (AI) and the prominence of large language models in mental health applications, questions persist about their aptitude in emotional comprehension. The prior iteration of the large language model from OpenAI, ChatGPT-3.5, demonstrated an advanced capacity to interpret emotions from textual data, surpassing human benchmarks. Given the introduction of ChatGPT-4, with its enhanced visual processing capabilities, and considering Google Bard's existing visual functionalities, a rigorous assessment of their proficiency in visual mentalizing is warranted. OBJECTIVE: The aim of the research was to critically evaluate the capabilities of ChatGPT-4 and Google Bard with regard to their competence in discerning visual mentalizing indicators as contrasted with their textual-based mentalizing abilities. METHODS: The Reading the Mind in the Eyes Test developed by Baron-Cohen and colleagues was used to assess the models' proficiency in interpreting visual emotional indicators. Simultaneously, the Levels of Emotional Awareness Scale was used to evaluate the large language models' aptitude in textual mentalizing. Collating data from both tests provided a holistic view of the mentalizing capabilities of ChatGPT-4 and Bard. RESULTS: ChatGPT-4, displaying a pronounced ability in emotion recognition, secured scores of 26 and 27 in 2 distinct evaluations, significantly deviating from a random response paradigm (P<.001). These scores align with established benchmarks from the broader human demographic. Notably, ChatGPT-4 exhibited consistent responses, with no discernible biases pertaining to the sex of the model or the nature of the emotion. In contrast, Google Bard's performance aligned with random response patterns, securing scores of 10 and 12 and rendering further detailed analysis redundant. In the domain of textual analysis, both ChatGPT and Bard surpassed established benchmarks from the general population, with their performances being remarkably congruent. CONCLUSIONS: ChatGPT-4 proved its efficacy in the domain of visual mentalizing, aligning closely with human performance standards. Although both models displayed commendable acumen in textual emotion interpretation, Bard's capabilities in visual emotion interpretation necessitate further scrutiny and potential refinement. This study stresses the criticality of ethical AI development for emotional recognition, highlighting the need for inclusive data, collaboration with patients and mental health experts, and stringent governmental oversight to ensure transparency and protect patient privacy.

The plasticity of ChatGPT's mentalizing abilities: personalization for personality structures.

This study evaluated the potential of ChatGPT, a large language model, to generate mentalizing-like abilities that are tailored to a specific personality structure and/or psychopathology. Mentalization is the ability to understand and interpret one's own and others' mental states, including thoughts, feelings, and intentions. Borderline Personality Disorder (BPD) and Schizoid Personality Disorder (SPD) are characterized by distinct patterns of emotional regulation. Individuals with BPD tend to experience intense and unstable emotions, while individuals with SPD tend to experience flattened or detached emotions. We used ChatGPT's free version 23.3 and assessed the extent to which its responses akin to emotional awareness (EA) were customized to the distinctive personality structure-character characterized by Borderline Personality Disorder (BPD) and Schizoid Personality Disorder (SPD), employing the Levels of Emotional Awareness Scale (LEAS). ChatGPT was able to accurately describe the emotional reactions of individuals with BPD as more intense, complex, and rich than those with SPD. This finding suggests that ChatGPT can generate mentalizing-like responses consistent with a range of psychopathologies in line with clinical and theoretical knowledge. However, the study also raises concerns regarding the potential for stigmas or biases related to mental diagnoses to impact the validity and usefulness of chatbot-based clinical interventions. We emphasize the need for the responsible development and deployment of chatbot-based interventions in mental health, which considers diverse theoretical frameworks.

ChatGPT outperforms humans in emotional awareness evaluations.

The artificial intelligence chatbot, ChatGPT, has gained widespread attention for its ability to perform natural language processing tasks and has the fastest-growing user base in history. Although ChatGPT has successfully generated theoretical information in multiple fields, its ability to identify and describe emotions is still unknown. Emotional awareness (EA), the ability to conceptualize one's own and others' emotions, is considered a transdiagnostic mechanism for psychopathology. This study utilized the Levels of Emotional Awareness Scale (LEAS) as an objective, performance-based test to analyze ChatGPT's responses to twenty scenarios and compared its EA performance with that of the general population norms, as reported by a previous study. A second examination was performed one month later to measure EA improvement over time. Finally, two independent licensed psychologists evaluated the fit-to-context of ChatGPT's EA responses. In the first examination, ChatGPT demonstrated significantly higher performance than the general population on all the LEAS scales (Z score = 2.84). In the second examination, ChatGPT's performance significantly improved, almost reaching the maximum possible LEAS score (Z score = 4.26). Its accuracy levels were also extremely high (9.7/10). The study demonstrated that ChatGPT can generate appropriate EA responses, and that its performance may improve significantly over time. The study has theoretical and clinical implications, as ChatGPT can be used as part of cognitive training for clinical populations with EA impairments. In addition, ChatGPT's EA-like abilities may facilitate psychiatric diagnosis and assessment and be used to enhance emotional language. Further research is warranted to better understand the potential benefits and risks of ChatGPT and refine it to promote mental health.

Lifestyle Changes, Emotional Eating, Gender, and Stress during COVID-19 Lockdown.

Emotional eating poses health risks. It is associated with adverse weight gain and a higher body mass index and is frequently triggered by stressful situations such as pandemics. The COVID-19 pandemic was found to cause stress as well as lifestyle changes of different magnitudes. This study examined the relationship between lifestyle changes and emotional eating by focusing on the moderating effect of gender and COVID-19-related stressors. A total of 1969 respondents completed online questionnaires during the first COVID-19 lockdown in Israel. The questionnaires assessed COVID-19-related lifestyle changes concerning eating habits, alcohol consumption, sleep quality, physical activity, and cigarette smoking, COVID-19-related stressors, and emotional eating (Dutch Eating Behavior Questionnaire). People with positive and negative COVID-19-related lifestyle changes demonstrated higher emotional eating compared to people with no lifestyle changes. The relationship between lifestyle changes and emotional eating was moderated by gender and COVID-19-related stressors. In conclusion, health officials should consider recommendations about lifestyle changes given to the public in times of unpredictable changes, with special attention to populations at risk of emotional eating. As emotional eating is related to poor emotional regulation skills, public nutrition recommendations should focus on improving these skills rather than adopting better diets.

Difference between Minorities and Majorities in the Association between COVID-19-Related Stress and Psychological Distress: A Socio-Ecological Perspective and the Moderating Role of Parenthood.

This study introduces a socio-ecological perspective of differences in psychological distress between the Palestinian minority and Jewish majority citizens of Israel during lockdown due to COVID-19. The study examines the association between COVID-19-related stress and psychological distress, and the moderating effect of parenthood. Online questionnaires, completed by 1934 participants (1391 Jews, 552 Palestinians; 1306 parents, 637 without children; 54.86% female, 45.13% male; M age = 40.38, SD = 13.77) assessed COVID-19-related stressors and depression, anxiety, and stress. The Palestinian minority showed a higher level of COVID-19-related stress and psychological distress than the Jewish majority. Parenthood showed a moderating effect on the association between COVID-19-related stress and distress for the Jewish majority but not the Palestinian minority. The results emphasize the significance of considering social status when seeking to understand the differences between minorities and majorities in terms of distress and resilience during pandemic events, and the need for cultural sensitivity and awareness when issuing instructions in such circumstances. Additionally, the results highlight the potential role of parenthood as a resilience factor, depending upon social status.

The Association Between Men's Mental Health During COVID-19 and Deterioration in Economic Status.

This study investigated associations among economic status deterioration, mental health, and gender during the COVID-19 pandemic. A total of 1,807 participants completed an online questionnaire that included demographic variables and questions measuring three mental health variables: psychological distress (as measured by symptoms of depression, anxiety, and stress), adjustment disorder, and emotional eating. Results indicated that women reported higher mental health impairment than men. Men and women whose economic status significantly deteriorated because of the COVID-19 pandemic reported greater mental health impairment than those whose economic status did not significantly deteriorate. However, men whose economic status significantly deteriorated reported high mental health impairment (emotional eating and adjustment difficulties) similar to women in the same situation. This change in men's reporting pattern suggests that the economic impact of COVID-19 severely impacted their mental health and affected how they view their masculinity, which, in turn, further impaired their mental health. As the COVID-19 outbreak has had a significant impact on mental health worldwide, it is important to identify individuals and groups who are at high risk of mental health impairment. The current study demonstrates that men's distress, which is frequently complex to identify, can be detected using standardized measures and analyzing these according to changes in reporting patterns as opposed to simply examining means and frequencies. The results suggest that the COVID-19 crisis may provide an opportunity to understand more about mental health, in particular, that of men.

The Association between Lifestyle Changes and Psychological Distress during COVID-19 Lockdown: The Moderating Role of COVID-Related Stressors.

The COVID-19 pandemic and the accompanying circumstances (lockdown and social distancing) have been found to cause lifestyle habit changes. While negative changes (adopting risky behaviors) are known to be associated with high psychological distress, the effect of positive changes (adopting health-enhancing behaviors) has yet to be investigated. This study examined the association between the nature of changes and psychological distress, in addition to the moderating effect of "COVID-related stressors". Online questionnaires, completed by 1969 participants, assessed the following: lifestyle changes due to COVID-19 with regard to cigarette smoking, alcohol consumption, physical activity, sleep quality, and eating habits; COVID-related stressors; Depression, Anxiety, and Stress Scale. Positive and negative changes were associated with higher psychological distress compared to no lifestyle change, and there was a moderating effect of COVID-related stressors in the association between COVID-related lifestyle changes and distress. In conclusion, to reduce psychological distress in overwhelming circumstances such as the COVID-19 pandemic, people should carefully consider whether to make changes in their lifestyle habits before doing so, even if these changes are perceived as positive and health-enhancing. Furthermore, efforts should be focused on reducing the perception of stress from COVID-19 by working on personal and mental perceptions of the situation.

Sleep quality and COVID-19-related stress in relation to mental health symptoms among Israeli and U.S. adults.

OBJECTIVES: The COVID-19 virus has resulted not only in high rates of morbidity and mortality across the globe, but in widespread mental health problems and sleep disruption, likely as a result of pandemic-related stressors. The current study examines associations among COVID-related stress, sleep quality, and mental health. DESIGN: Cross-sectional data were collected via online surveys in May 2020. PARTICIPANTS: were 2541 community adults ages 18-70 from Israel (N = 1969) and the U.S. (N = 572). MEASUREMENT: Participants completed measures of COVID-related stress, sleep quality, and symptoms of anxiety, depression and adjustment disorder. RESULTS: Participants reported high rates of depression and anxiety symptoms, adjustment difficulties, and poor sleep quality. In both countries, COVID-related stressors were associated with both anxiety and depression, and these associations were mediated by sleep disturbances. CONCLUSIONS: These results support the role of sleep in mental health difficulties. Widespread, accessible, evidence-based interventions are urgently needed to improve health and mental health and to promote resilience in preparation for future global crises.

Social relations between students from two groups in conflict: Differences in stereotypes and perceived social distance between Jewish and Arab nursing students.

BACKGROUND: The heterogeneous nature of the nursing profession generates encounters between members of different groups. Thus, nursing education should include programs focusing on cultural sensitivity to enhance cultural competence. Due to the ongoing conflict between Arabs and Jews, Israel is an appropriate setting for examining cultural sensitivity programs. AIM: This study explores how Jewish-Israeli and Arab-Israeli nursing students studying together perceive their social relations by examining attribution of stereotypes (positive and negative) to members of the other group and willingness to engage in social interactions with individuals from the other group. SETTING: The Department of Nursing at the Max Stern Yezreel Valley College. PARTICIPANTS: 180 Jewish and Arab nursing students at two different points in their studies, beginning students and those at more advanced stages. METHODS: The students completed questionnaires examining their perceptions of the social relations between the two groups. The questionnaires measured stereotypes and perceptions of social distance regarding members of the other social group. RESULTS: More positive stereotypes and fewer negative stereotypes were found among the Jewish students than among the Arab students. Nevertheless, the Jewish students exhibited less willingness to reduce social distance. No differences were found between beginning students and advanced students with respect to stereotypes, though advanced students expressed greater willingness to reduce social distance than did beginning students. CONCLUSIONS: The findings show that perceptions of social relations are influenced by two different contexts: the students' specific context (joint studies and joint work in the future) and the broader general context of relations between the groups. The findings indicate that programs aimed at developing cultural sensitivity and improving relations in a divided society should provide differential responses for each group and should focus on willingness to cooperate and reduce social distance rather than attempting to eliminate stereotypes.