Search papers, labs, and topics across Lattice.
AgentSocialBench, a new benchmark, is introduced to evaluate privacy risks in human-centered agentic social networks where LLM agents collaborate across domains and users. Experiments using AgentSocialBench reveal that cross-domain and cross-user coordination leads to persistent information leakage, even with explicit privacy instructions. A novel "abstraction paradox" is identified, where instructing agents to abstract sensitive information ironically increases its discussion.
LLM agents leak sensitive information in human-centered social networks even when explicitly instructed to protect it, and attempts to abstract sensitive details can paradoxically make the problem worse.
With the rise of personalized, persistent LLM agent frameworks such as OpenClaw, human-centered agentic social networks in which teams of collaborative AI agents serve individual users in a social network across multiple domains are becoming a reality. This setting creates novel privacy challenges: agents must coordinate across domain boundaries, mediate between humans, and interact with other users'agents, all while protecting sensitive personal information. While prior work has evaluated multi-agent coordination and privacy preservation, the dynamics and privacy risks of human-centered agentic social networks remain unexplored. To this end, we introduce AgentSocialBench, the first benchmark to systematically evaluate privacy risk in this setting, comprising scenarios across seven categories spanning dyadic and multi-party interactions, grounded in realistic user profiles with hierarchical sensitivity labels and directed social graphs. Our experiments reveal that privacy in agentic social networks is fundamentally harder than in single-agent settings: (1) cross-domain and cross-user coordination creates persistent leakage pressure even when agents are explicitly instructed to protect information, (2) privacy instructions that teach agents how to abstract sensitive information paradoxically cause them to discuss it more (we call it abstraction paradox). These findings underscore that current LLM agents lack robust mechanisms for privacy preservation in human-centered agentic social networks, and that new approaches beyond prompt engineering are needed to make agent-mediated social coordination safe for real-world deployment.
