Unverified50% confidenceFactExact time
Academic researchers and platform security researchers have documented multiple cases of using large language models to batch-generate high-quality community comments to manipulate public opinion, with some research referring to this as LLM-driven astroturfing.
1
Sources
50%
Confidence
Long-term
Relevance
9/25/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedMIT媒体实验室、斯坦福HAI研究所等机构正在系统性研究人机混合社区中的信息传播、信任机制和群体决策动态65% similarUnverifiedConversations with large language models can substantially influence users' attitudes and beliefs, spanning areas from product recommendations to political positions.64% similarUnverifiedModel guardrail bypass methods documented by the security research community include role-play prompt injection, multi-step decomposed requests, and adversarial prefixes64% similarUnverifiedA 2016 fMRI study by Jonas Kaplan et al. at the University of Southern California showed that participants' default mode network activity significantly increased when their political beliefs were challenged63% similarUnverified对初级分析师而言，将脱敏的真实项目拿到数据社区请教，比抽象提问能获得更有深度的反馈63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/950760API
curl https://kongchang.com/api/v1/knowledge/claims/950760MCP
get_claim(id=950760)