A new position paper published on arXiv outlines the risks associated with Generative Engine Optimization (GEO), a phenomenon where search engine optimization (SEO) techniques are adapted for large language model (LLM) answer engines. The paper identifies three key risks: concentrated influence due to low contestability and system sensitivity, undisclosed commercial influence within LLM-generated answers, and blind spots arising from asymmetries between academic evaluations and real-world deployed systems. To address these issues, the authors advocate for answer-level governance, including enhanced contestability, precise disclosure of influence, black-box auditing, and deployment-aligned metrics. AI
IMPACT Highlights potential manipulation of LLM-based information sources, necessitating new governance and auditing approaches.
RANK_REASON The cluster contains an academic paper discussing risks and proposing governance for a new AI-related phenomenon. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →