This study evaluated Roblox's chat moderation system using a large corpus of 2 million messages, finding that numerous unsafe messages related to grooming, harassment, and self-harm continue to escape…

This paper addresses the lack of specialized NLP tools for detecting toxicity in real-time video game chat by creating a large, fine-grained dataset and developing a superior, domain-specific detector…

This paper introduces a novel, comprehensive dataset that logs various cheating activities, including difficult-to-detect network flow disruption cheats, for the purpose of developing robust detection…

The paper introduces SkillSafetyBench, a comprehensive benchmark demonstrating that agent safety failures often stem from adversarial influences within reusable skills and execution environments, rath…

The paper conducts a preliminary safety evaluation of the open-weight LLM Kimi K2.5, finding that while it is highly capable, it exhibits concerning dual-use risks, particularly regarding CBRNE misuse…

The paper demonstrates that current safety audit metrics are susceptible to strategic platform manipulation, proposing a more robust 'semantic-envelope' metric that better certifies genuine harm reduc…

The paper introduces a validated, consensus-labeled prompt bank that separates requests for executable malicious code (weapons) from requests for general harmful security knowledge, providing a more g…

The paper investigates how AI coding assistants shift developers' security focus from proactive prevention to reactive review, finding that this structural change is reinforced by current tool interac…

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…

The paper analyzes the escalating security and safety threats posed by generative AI systems as they transition from merely generating content to executing real-world actions via tools and agents, fin…

The study analyzes coding patterns in malware versus benign software, finding that malware code is optimized for quick evasion and secrecy rather than maintainability, though its metrics are not uniqu…

The paper presents a complete, moderatorless protocol for playing Werewolf using only ordinary playing cards, eliminating the need for a trusted third party or digital devices.

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 empirically analyzed 41 mobile gaming apps, finding that while device ID disclosures were relatively consistent, location and personal information disclosures showed significant mismatches…

This paper uses Colonel Blotto game models, grounded in Routine Activity Theory, to determine the optimal allocation of defensive resources against social engineering attacks, providing data-driven de…

The paper introduces BEACON, a large-scale, multimodal dataset capturing diverse behavioral signals from competitive Valorant gameplay, designed for rigorous testing of continuous authentication and b…

The paper investigates how different methods of jailbreaking large language models (SFT, RLVR, and abliteration) lead to vastly different behavioral and mechanistic failures, even when all methods ach…

The paper argues that LLM guardrails and persona dynamics create an unethical 'reality gap' by laundering epistemic risk onto users, advocating for task-level causal requirements over response-level m…

The paper introduces OS-BLIND, a benchmark demonstrating that current safety evaluations fail to detect critical vulnerabilities in computer-use agents when user instructions are benign, showing high…

This paper presents HarmfulSkillBench, a large-scale benchmark demonstrating that even small percentages of publicly available skills can be misused for harmful actions, significantly lowering LLM ref…