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.

Quantifying the Predictive Role of Temperament Dimensions and Attachment Styles on the Five Factor Model of Personality.

BACKGROUND: The present study investigated the role of temperament and attachment security in predicting individual differences in the five factor personality traits among adults. As previous studies suggested the potential moderating role of attachment in the association between temperament and personality traits, the present study sought to examine an interactionist model combining attachment and temperament in explaining individual differences in personality traits. METHODS: A sample of 1871 participants (1151 women and 719 men) completed self-report measures of adult attachment style (the Relationships Questionnaire-RQ), temperament dimension (the Fisher Temperament Inventory-FTI), and personality domain (the Five Factor Model-FFM). RESULTS: Partial correlational analyses revealed associations between attachment security and each of the five domains of the FFM, and few associations between some temperament dimensions and several domains of the FFM. Moderated regression analyses showed that attachment security moderated the associations between temperament dimensions and the Agreeableness domain of the FFM. Among secure individuals, those with higher scores on the Curious/Energetic, Cautious/Social Norm Compliant and Prosocial/Empathetic scales exhibited higher Agreeableness scores, whereas among insecure individuals, those with higher scores on the Analytic/Tough-minded scale exhibited lower scores on the Agreeableness scale. CONCLUSION: Overall, the current study provides evidence in support of the substantive role of social-environmental factors (Adult Attachment) as a moderating element bridging temperament-related personality elements and a number of their FFM manifestations.

Factors affecting compliance with use of online healthcare services among adults in Israel.

BACKGROUND: The use of online health services (henceforth, OHS) among middle-aged to older adults can make health-related actions more accessible to this population group as well as help reduce the burden on the health system and avoid unnecessary costs. The study's objectives were to examine the responsiveness and willingness of individuals aged 45+ to use different OHS and to characterize the attitudes and main factors influencing that responsiveness. METHODS: We conducted a telephone survey among a sample of 703 individuals constituting a representative sample of the Israeli population of individuals aged 45+. The research questionnaire integrates the principles of the Adopting Medical Information Technologies model and includes socio-demographic attributes. RESULTS: The results indicate that 78 % of internet users claimed to use at least one OHS (79 % of the Jewish sector and 66 % of the non-Jewish sector). Nevertheless, 22 % of internet users do not use OHS. Most online use is on Health Maintenance Organizations (HMO) websites to obtain administrative information. Frequency of OHS use increases as the following factors increase: perceived ease of online use; extent of encouragement for online use; perceived reliability of online health services; and extent of advertisement exposure. The study found that OHS use is much more prevalent among wealthy populations. In addition, individuals' attitudes and the extent of their exposure to advertisement influence their use and intention to use OHS. CONCLUSIONS: A number of recommendations emerge from the study: 1) For OHS use to increase online health websites should be made more accessible to middle aged-older adults individuals and those of different languages and cultures. 2) Programs should be developed to teach HMO staff to encourage patients to use OHS. 3) Media advertising that encourages the use of OHS should be expanded.