Researchers Identify Governance Gaps in Generative Engine Optimization as LLMs Reshape Search
A new position paper accepted to ICML 2026 argues that Generative Engine Optimization (GEO)—the practice of optimizing content for LLM-based answer engines—creates three underexamined risks: concentrated influence, undisclosed commercial bias, and blind spots between academic research and deployed systems. The shift from traditional search engine rankings to synthesized LLM answers changes how information visibility works online. The authors call for stronger governance including disclosure requirements, contestability mechanisms, and auditing standards aligned with real-world deployment.
Researchers have identified significant governance gaps emerging as large language models increasingly serve as answer engines for information seeking, replacing traditional ranked search results with synthesized responses. This shift enables Generative Engine Optimization (GEO), a practice analogous to SEO but targeting the evidence pools and generation processes of LLM systems. The paper formalizes a GEO pipeline and identifies three key risks: concentrated influence from low contestability and system sensitivity, undisclosed commercial influence embedded in evidence and reasoning, and blind spots between academic evaluation practices and deployed systems. The authors argue these gaps stem from visibility and evaluation asymmetries between offline research setups and real-world deployments. They propose targeted governance measures including answer-level accountability, high-precision disclosure of commercial influence, black-box auditing of material influence, and metrics that reflect actual exposure persistence in deployed systems.
What's missing
The paper does not provide empirical data on the prevalence or scale of GEO practices in current deployed systems, nor does it quantify the magnitude of influence concentration or commercial bias in existing LLM answer engines. The specific mechanisms by which academic and industry evaluation practices diverge are not detailed in the abstract.
What different sources said
- arXiv cs.AICenter
Position: Generative Engine Optimization Creates Underexamined Risks, Governance Must Target Concentration, Disclosure, and Academic Blind Spots
Related
Topology-Aware Thermodynamics Improves DNA Probe Specificity Design
Researchers developed a new framework for designing DNA probes that accounts for the spatial organization of matched sequences, not just overall thermodynamic stability. Traditional methods rely on scalar measures like melting temperature and free energy, which miss how mismatches are distributed along the probe. The approach could improve diagnostic accuracy in applications like HPV detection and gene expression profiling.
Study Identifies Optimal Thermal Dose for Combining Focused Ultrasound with Immunotherapy in Tumors
Researchers used multimodal PET imaging to identify an optimal thermal dose range for focused ultrasound ablation that destroys tumor tissue while preserving conditions for immunotherapy delivery. The study found that excessive heating collapses blood vessels needed for antibody access, while insufficient heating fails to adequately reduce tumor burden. The findings could guide clinical design of combination treatments pairing thermal ablation with immunotherapies.
Plant MSH1 Protein Functions as Mismatch-Directed Nuclease for Organelle Genome Maintenance
Researchers have identified the precise mechanism by which the AtMSH1 protein in Arabidopsis plants recognizes and cleaves DNA mismatches and lesions, preventing mutations in organellar genomes. The protein combines a DNA mismatch recognition module with a nuclease domain that makes staggered cuts at specific positions relative to DNA damage. This discovery explains how plants maintain unusually low mutation rates in their mitochondrial and chloroplast DNA compared to other eukaryotes.