SOURCE-LINKED INTELLIGENCE

Evaluating Deep-Search Agents under Hierarchical Web Evidence Poisoning

arXiv · AI, language, vision and robotics · article · Sep 5, 2026 · UTC

Search-augmented LLM agents are increasingly used for consumer decisions, making them vulnerable to Generative Engine Optimization (GEO) poisoning. Existing benchmarks largely measure whether manipulated content is retrieved or endorsed, but do not track whether an agent verifies suspicious evidence, revises adopted claims, or recovers before producing its final recommendation. We introduce HAE-GEO, a benchmark that tracks the full trajectory from exposure to recovery under progressively more persuasive Web poisoning. Agents interact via a multi-turn Search-Scrape interface across three attack

Read original source ↗ Open in workspace

recordType

paper

region

Global