PhD Defense: Desheng Hu on the Societal Impact of Web Search

Search engines play a central role in how people access information online. As Google increasingly integrates generative AI features such as AI Overviews, questions about the influence, quality, and safety of search results are becoming even more important.

In his PhD thesis, Desheng Hu examined web search through the lens of Google Search, a large-scale algorithmic system with far-reaching societal implications. Combining the analysis of large-scale observational data with algorithm auditing experiments, he investigated three key areas: Google's market dominance, the amplification of political partisanship in search snippets, and the quality and safety of AI-generated health information.

His research also developed an evaluation framework and annotation guidelines for assessing AI-generated health information. These tools examine factors such as consistency, relevance, source credibility, and medical safeguards, helping to identify potential risks to users' well-being.

The findings contribute to discussions about competition in digital markets, exposure to political information, and the trustworthiness of AI-generated content. They also underline the importance of independent evaluations to support safer and more responsible AI systems.

Desheng Hu conducted his PhD research in the Social Computing Group (SCG) at the Department of Informatics under the supervision of Prof. Dr. Anikó Hannák, with Prof. Dr. Christo Wilson (Northeastern University) as co-advisor.