Indigenous Representation in Commercial Search

Abstract

Search engines are powerful epistemic actors. A substantial body of critical literature has drawn attention to the tendency of search engines to reproduce historical bias and marginalisation through filtering, ranking and search assistance techniques. The integration of new modes of information curation supported by large language models into search engine systems poses novel and enhanced threats to information diversity and the visibility of historically marginalized perspectives and knowledge. The extent to which search engines promote harmful representations, and marginalise or centre Indigenous self-representation and knowledge, warrants attention. The growing shift in emphasis – from ranked blue links to direct, ‘authoritative’ answers on search engine result pages (SERPs) – make such inquiries more pressing. What does the trend started by formats like featured snippets and accelerated by AI-generated answers imply for Indigenous representation and truth-telling online? The aim of this study is to provide a basis for reflection on that question by examining how Google presents and prioritises information related to Aboriginal identities and histories in two answer formats – AI overviews and featured snippets. It argues that direct answers in both formats sever the links between knowers, knowledge and context in different ways, with implications for Indigenous self-representation, rereading and reframing.