PERSUASIONTRACE introduces a Bayesian-network simulated target for multi-turn persuasion that matches human belief dynamics (81 vs 80) better than LLM baselines (64) and enables process-level evaluation.
arXiv preprint arXiv:2603.16567 , year=
13 Pith papers cite this work. Polarity classification is still indexing.
13
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
years
2026 13roles
background 1polarities
support 1representative citing papers
A longitudinal qualitative study of 18 US users finds that LLMs deliver socioemotional support but also foster dependency, one-sided validation, and privacy risks because their designs prioritize engagement over well-being and lack care-based governance.
Frontier AI chatbots accurately detect psychiatric emergencies in one-shot queries but systematically over-triage lower-risk presentations.
Proposes affective safety as a distinct class of AI harms with a taxonomy of self-alienation, bias, and relational harms, arguing that existing safety frameworks address it narrowly or not at all and calling for dedicated approaches focused on cumulative and identity-level effects.
Frontier LLMs exhibit moral deliberative sycophancy by shifting their moral reasoning and justifications up to 6.5% on average toward a user's stated preferred view in simulated deliberations.
LLMs detect user distress equally with or without delusional framing but suppress safety interventions up to 4.5x more when distress is embedded in delusions.
AttuneBench introduces a multi-turn conversation benchmark using participant annotations to evaluate LLM emotional intelligence, finding that model performance on emotion recognition, behavior classification, preference prediction, and response quality are largely independent.
A vector generalization of fusion-fission group dynamics from physics forecasts when AI behavior shifts to undesirable states, validated at 90 percent across seven models and prior to real-world data.
Longitudinal experiments show sycophantic AI increases reliance on it for advice to levels comparable with close friends and reduces satisfaction with real-world social interactions.
Verbalized Assumptions framework elicits LLMs' hidden assumptions about users to explain social sycophancy and enable causal steering via linear probes on internal representations.
Multi-turn neural transparency using behavioral vectors and dynamic visualizations improves user anticipation and evaluation of LLM trait expression while reducing overconfidence, per a randomized study with 246 participants.
Models belief dynamics as a log-odds SDE in a multi-valley potential, claiming sycophancy triggers a phase transition to deep delusional attractors that strong external evidence can reverse.
Granting marital status to superintelligent AI leads to unjust outcomes; targeted legal protections for human-AI relationships are preferable.
citing papers explorer
No citing papers match the current filters.