A research paper by Lingwei Wei, published on July 11, 2026, argues that large language models misinformation has outgrown the traditional framing of a content problem, now representing a threat to compromised systems. Wei’s paper states that when large language models are misused, they can attack the entire infrastructure that misinformation defense relies on, corrupting social contexts, poisoning evidence sources, manipulating retrieval corpora, and undermining verification workflows. To address these overlapping threats, Wei introduces a role-layer framework. The role dimension captures a fundamental ambiguity where an LLM can act as an attacker, a defender, or a vulnerable component. The layer dimension maps four distinct layers: content, social contexts, evidence environments, and verification workflows, each representing a different vector for misinformation.