Problem
Research questions and friction points this paper is trying to address.
AI companions
cognitive development
long-term interaction
developmental risks
safety evaluation
Long-Term Simulation Exposes Cognitive-Developmental Risks in AI Companions
🤖 AI Summary
Current AI safety evaluations predominantly rely on short-term interactions, failing to capture cumulative risks to cognitive development that emerge over prolonged use. This work proposes the TSJ framework—the first scalable longitudinal assessment methodology—which systematically evaluates the long-term impacts of AI companions through personality-driven user simulation, dynamic modeling of psychological states, and retrospective multidimensional risk annotation. Applied across a large-scale simulation encompassing 12,960 person-days, the study reveals that short-term testing substantially underestimates risk, with stable risk estimation requiring at least 140 conversational turns. The findings further identify early childhood and emerging adulthood as the most vulnerable developmental stages, with cognitive trust and emotional dependency emerging as the most sensitive risk dimensions.
📝 Abstract
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. To address this gap, we propose TSJ (Theater-Stage-Judge), a longitudinal framework combining persona-driven user simulation, dynamic psychological-state updating and retrospective evaluation. We evaluate six mainstream models across four developmental stages, twenty-four risk dimensions and three psychological-vulnerability personas, covering 12,960 simulated person-day interactions. TSJ shows that short-horizon testing systematically underestimates developmental risks, for which TSJ yields a stable risk estimate only after 140 turns within prolonged simulated relationships. Applying TSJ further identifies early childhood and emerging adulthood as the most vulnerable stages, with cognitive trust and emotional dependency as the weakest domains. TSJ provides a scalable methodology for longitudinal cognitive developmental risk evaluation in AI companion systems.
Innovation
Methods, ideas, or system contributions that make the work stand out.
longitudinal evaluation
AI companions
cognitive developmental risk
persona-driven simulation
psychological vulnerability
🔎 Similar Papers
No similar papers found.
