arXiv:2608. 02491v2 Announce Type: replace Abstract: Language models have taken on the role of a very new type of technology, by virtue of their "human-ness" and rapid integration into users' daily lives.
arXiv:2601. 07988v2 Announce Type: replace-cross Abstract: While NLP typically treats documents as independent and unordered samples, in longitudinal studies, this assumption rarely holds: documents are nested within authors and ordered in time, forming person-indexed, time-ordered $\textit{behavioral sequences}$.
AI companions powered by large language models increasingly interact with cognition-developing users, including children and adolescents, creating risks that may accumulate over time. Existing safety evaluations largely rely on single-turn or short-session tests, which cannot capture risks that emerge only through prolonged interaction.
arXiv:2607. 26820v1 Announce Type: new Abstract: As large language models (LLMs) evolve from standalone assistants into autonomous agents, ensuring their safety requires shifting beyond pointwise risk assessment to understand how risks emerge and unfold over long-horizon trajectories.
The article surveys how large language models (LLMs) are being applied to mental health, outlining a three‑phase evolution: Phase I uses LLMs as passive information tools and pattern recognizers for assessment; Phase II employs them as empathetic conversationalists for stateless, in‑the‑moment interactions; Phase III aims to create longitudinal, personalized companions that act as stateful cognitive agents. It systematically reviews core technologies, agent architectures (Profile, Memory, Reasoning, Planning), and the datasets and benchmarks that support this progression, offering a coherent narrative and roadmap for future research. The survey also provides a curated resource list at https://github.com/Emo-gml/Awesome-Mental-Health-LLMs.
arXiv:2606. 19144v1 Announce Type: new Abstract: Current conversational AI systems have made significant progress in language generation, personalization, and long-context interaction.
arXiv:2606. 29685v1 Announce Type: new Abstract: How can we evaluate whether frontier AI systems recognize child-safety risks before they escalate into explicit harm?
arXiv:2607. 19361v1 Announce Type: cross Abstract: Most safety guardrails for large language models (LLMs) evaluate each prompt-response pair in isolation, which misses failures that arise only over a dialogue as benign turns compose into harm.
The paper examines how changes made by platforms can disrupt users’ relationships with AI companions. It catalogs 30 disruption events, creates a taxonomy of six types, and identifies three main causes. A risk‑assessment framework with four dimensions is proposed, and a Bayesian time‑series analysis of Reddit data shows that disruptions trigger spikes in anxiety, stress, suicidal expression, and grief, especially when relational continuity and transition support are lacking.
The paper introduces the Core Sentiment Inventory (CSI), a new personality trait evaluation tool for large language models (LLMs) that addresses reliability and validity issues found in existing methods like the Big Five Inventory (BFI). CSI is designed specifically for LLMs, supports both English and Chinese, and provides detailed psychological portraits of model behavior. Experiments show that CSI captures nuanced behavioral patterns, improves reliability, and correlates strongly (above 0.85) with real-world LLM outputs.
arXiv:2606. 08945v1 Announce Type: new Abstract: We investigate whether information about time-to-event risk estimated by a Cox proportional hazards model can be transferred into a generative large language model.
arXiv:2606. 31464v1 Announce Type: cross Abstract: Recent advances in Large Language Models (LLMs) have motivated their adoption across a wide range of domains, including Artificial Intelligence (AI) for mental health.
CHRONOBERG is a temporally structured corpus of English book texts covering 250 years, curated from Project Gutenberg and enriched with temporal annotations. It enables quantification of lexical semantic change via time‑sensitive Valence‑Arousal‑Dominance analysis and the creation of historically calibrated affective lexicons. Experiments show that language models trained sequentially on CHRONOBERG struggle to encode diachronic shifts, highlighting the need for temporally aware training and evaluation pipelines.