Ali Sahandi | Searcharxiv

arXiv2026

Self-Expansion Through AI: Latent Profile Analysis of AI Companionship and Wellbeing Outcomes

People increasingly form meaningful relationships with conversational AI. Current variable-centered literature offers a mixed picture of relational patterns and how they associate with wellbeing. Taking a person-centered approach, we applied latent profile analysis to secondary data from 467 active conversational AI users, evaluating five indicators of machine companionship: utilitarian motivation, eudaimonic exchange, connective coordination, mind perception, and inclusion of other in the self. Five profiles emerged, ordering parsimoniously from objectified-utilitarian to intimate-bonded. Conceptually divergent utilitarian motivation and companionate engagement co-occurred, rather than diverging. Furthermore, deep companionship was achieved without users attributing a mind to the AI. Profile membership was strongly associated with relatedness need satisfaction, tracking upward across the groups, but was only weakly associated with loneliness, indicating AI fulfills immediate relational needs but may not resolve systemic human disconnection. Ultimately, these findings characterize human-AI relationships as graded configurations spanning a superordinate continuum of self-expansion

cs.HC↗

arXiv2026

WET -- Weighted Ensemble Transformer for Identifying Psychiatric Stressors Related to Suicide on X (formerly Twitter)

Suicide remains one of the leading causes of death worldwide, particularly among young people, and psychological stressors are consistently identified as proximal drivers of suicidal ideation and behavior. In recent years, social media platforms such as X have become critical environments where individuals openly disclose emotional distress and conditions associated with suicidality, creating new opportunities for early detection and intervention. Existing approaches, however, predominantly rely on raw textual content and often neglect auxiliary emotional and contextual signals embedded in user metadata. To address this limitation, we propose a Weighted Ensemble Transformer (WET), a dual branch deep learning architecture designed to identify psychiatric stressors associated with suicide in X posts. Our model integrates semantic representations extracted through Transformer encoders with an engineered feature vector capturing sentiment, subjectivity, polarity, and user engagement characteristics. We collected, filtered, and annotated 125,754 English tweets for suicide-related psychological stressors and evaluated the proposed model under two configurations. Extensive comparative experiments against traditional machine learning methods, advanced recurrent networks, and transformer baselines demonstrate that WET achieves state-of-the-art performance, reaching 0.9901 accuracy in binary classification. These findings show that hybridizing deep semantic signals with auxiliary emotional and behavioral features substantially improves suicidality detection accuracy.

cs.SI↗