Zinan Zhang

Zinan Zhang, Xinning Gui, Yubo Kou

Cooperative play (co-play) is often positioned as a family-beneficial practice that can strengthen parent-child bonds and support parental mediation in games. Yet co-play in user-generated virtual worlds (UGVWs) can be disrupted by real-time harms that parents cannot easily prevent. Roblox, a platform with millions of user-generated virtual worlds and a large child player base, illustrates this challenge. Prior work on harmful UGVW design highlights risks beyond content problems, including manipulative monetization prompts, unmoderated social interactions, emergent in-world behaviors, and narrative designs that may normalize harmful ideologies. Current governance and moderation approaches, largely adapted from social media, focus on static artifacts and often fail to capture interactive and emergent harms in virtual worlds. This workshop paper asks: how might UGVWs and their platforms be designed to minimize harms that specifically impair family co-play experiences?

Zinan Zhang, Xinning Gui, Yubo Kou

Emerging extended reality technologies are reshaping how children play, learn, and socialize. Yet, they also present serious safety risks. Gaming, a primary form of entertainment for children, is also one of the key applications of XR. While XR platforms offer immersive and engaging gaming experiences, recent news has highlighted safety concerns such as car accidents, lower judgment for real-world situations, and exposure to disturbing content like virtual rape. This research examines how XR game design may lead to online safety risks for children. Through analysis of player forums, game developer forums, and interviews with child players, we identify harmful XR design patterns, explore how developers collaboratively generate and implement risky game ideas, and document children's firsthand experiences of online safety risks. Existing ethical frameworks often fail to address the immersive and socially dynamic nature of XR games. We advocate for a child-centered, design-aware approach to ethical considerations in XR games, urging platforms and policymakers to prioritize children's developmental needs. Our work aims to help shape safer, more inclusive XR environments through research and cross-sector collaboration.

Zhe Yang, Shulin Tian, Kairui Hu et al.

We present HippoCamp, a new benchmark designed to evaluate agents' capabilities on multimodal file management. Unlike existing agent benchmarks that focus on tasks like web interaction, tool use, or software automation in generic settings, HippoCamp evaluates agents in user-centric environments to model individual user profiles and search massive personal files for context-aware reasoning. Our benchmark instantiates device-scale file systems over real-world profiles spanning diverse modalities, comprising 42.4 GB of data across over 2K real-world files. Building upon the raw files, we construct 581 QA pairs to assess agents' capabilities in search, evidence perception, and multi-step reasoning. To facilitate fine-grained analysis, we provide 46.1K densely annotated structured trajectories for step-wise failure diagnosis. We evaluate a wide range of state-of-the-art multimodal large language models (MLLMs) and agentic methods on HippoCamp. Our comprehensive experiments reveal a significant performance gap: even the most advanced commercial models achieve only 48.3% accuracy in user profiling, struggling particularly with long-horizon retrieval and cross-modal reasoning within dense personal file systems. Furthermore, our step-wise failure diagnosis identifies multimodal perception and evidence grounding as the primary bottlenecks. Ultimately, HippoCamp exposes the critical limitations of current agents in realistic, user-centric environments and provides a robust foundation for developing next-generation personal AI assistants.