Shi Young-hoo

Overview

Shi Young-hoo is the Arizona State University researcher interviewed in How easy is it to game AI search results? about manipulation of AI-search rankings. The name and affiliation are retained as supplied by the episode note.

Current Profile

Shi’s source-bounded role is to distinguish technically neutral ranking optimization from malicious promotion of fabricated or untrustworthy material. Shi describes text and image attacks, transfer from open-source models to proprietary systems, privacy-oriented demotion, and the need for users to inspect sources rather than accept a chatbot’s polished answer as authority.

Key Characteristics

• Studies or explains how content changes can alter ranking in LLM-based retrieval systems.
• Treats optimization techniques as dual use: they can promote scams or reduce exposure of personal information.
• Reports transfer effects across open-source and proprietary model families.
• Emphasizes source inspection and cross-system checking as practical user safeguards.

Evidence

• Ranking manipulation - How easy is it to game AI search results? has Shi discuss fabricated products, user-generated promotional content, model-preferred writing styles, and image perturbations.
• Cross-model transfer - How easy is it to game AI search results? attributes observed effects on GPT, Claude, and Gemini systems to techniques developed with open-source models.
• Dual use and verification - How easy is it to game AI search results? connects ranking demotion to privacy and recommends reviewing sources or using multiple models or agents to cross-check references.

Qualifications

The short episode does not identify the underlying paper, experimental protocol, model versions, quantitative results, or replication status. The spelling of Shi’s name, institutional role, and all research findings therefore remain tied to the supplied source rather than independently verified here.

What Changed

• Created a source-bounded profile for the researcher explaining AI-search ranking manipulation and verification.

Relationships

• AI Search Ranking Manipulation - research topic Shi explains through text, image, and cross-model attack examples.
• Answer Engine Optimization - legitimate optimization category whose deceptive boundary Shi helps define.
• AI Answer Evaluation - user-side source inspection and cross-system checking practice Shi recommends.
• Marketplace Tech - program that interviewed Shi.