This study explores why and how users customize social AI, identifying motivations and contributions to a closer relationship.

The study finds exploratory evidence that gender moderates how youth perceive privacy risks and benefits, influencing their protective behavior when using smart voice assistants.

The paper introduces EUDAIMONIA, a new framework and benchmark for evaluating how well LLMs align with user welfare in social interactions, finding that even state-of-the-art models frequently violate…

This study proposes a negotiation framework, using composite indices (RBTI and CATI), to explain how youth navigate competing privacy pressures when using smart voice assistants, finding that high usa…

This paper investigates how socially calibrated large language models distinguish when to incorporate others' perspectives from maintaining their own moral judgments, using three studies.

The paper introduces PHTV-Scout, a novel framework that analyzes Douyin and Kwai data, revealing a high prevalence of potentially harmful teen videos, particularly CSE imagery, and demonstrating that…

The study found that while contextualizing AI responses reduces their persuasive power, combining this technique with conversational warmth restores persuasiveness, suggesting that user deference to A…

The paper proposes AI From the Margins (AIM), a methodological stance that centers the lived experiences of minoritized communities to fundamentally reshape the goals and scope of participatory AI des…

Two student cohorts designed cognitively accessible GenAI interfaces, leading to structural and experiential scaffolding concepts.

The paper argues that LLM agent security is fundamentally an agent-human interaction (AHI) problem, demonstrating that industry practices rely on human-centric mechanisms while academic research focus…

This study surveyed higher education practitioners to map their beliefs and behaviors regarding AI integration, finding that while they view AI favorably, institutional barriers and gaps in design-ori…

This study investigated user reactions to inferred personal information from their own ChatGPT histories, finding that acceptability is governed by context-sensitive norms regarding generation, retent…

The study demonstrates that conditioning AI brand recommendations on a user's persona significantly alters the recommended product set, particularly for mid-market brands, and this effect is largest o…

This paper conducted a randomized controlled trial on Reddit to test the effectiveness of various deescalation strategies in reducing personal insults using automated replies.

The paper proposes a hierarchical framework, PHF (Practice-Habitus-Field), inspired by Bourdieu's Theory of Practice, to improve LLM personalization by modeling user behaviors at three distinct levels…

This paper identifies and validates eight core UX principles for human-AI agent interaction in the workplace using a multi-method approach.

The paper argues that traditional identity-based reputation mechanisms are structurally inapplicable to language model agents because their mutable, modular nature makes them ontologically dissociativ…

BuddyBench introduces a novel, privacy-constrained multi-task benchmark that integrates longitudinal learning trajectories, standardized clinical assessments, and randomized trial data to advance pedi…

The paper investigates how users perceive and utilize security and privacy transparency in consumer-facing generative AI, finding that users rely on proxies like popularity and require actionable, tru…

The paper investigates how digital devices in U.S. prisons create privacy and security risks for incarcerated users, finding that pervasive surveillance and arbitrary policies negatively impact their…