MAD-Spear: a Conformity-Driven Prompt Injection Attack on Multi-Agent Debate Systems

Multi-agent debate MAD systems leverage collaborative interactions among large language models LLMs agents to improve reasoning capabilities. While recent studies have focused on increasing the accuracy and scalability of MAD systems, their security vulnerabilities have received limited attention...

Election Security: How Mobile Devices Are Shaping the Way We Work, Play and Vote

The line between our personal and professional lives is blurring in an unprecedented fashion as we approach the 2020 presidential election. From Oracle and Walmart’s plans to invest in TikTok to a bug in Joe Biden’s campaign app that exposed millions of voter files – the role mobile technology wi...