AI-Generated Debates Between Public Thought Leaders on Masculinity
This system stages structured debates between AI agents representing 14 public thought leaders. Each agent draws from the speaker's actual YouTube transcripts using retrieval-augmented generation (RAG), filtered for faithfulness to their real positions. Debates follow a 3-round structure: foundational beliefs, challenges and provocations, then rhetorical evidence.
Select a research question below to explore the debates, including moderator analysis and synthesis of each exchange.
The debate transcripts on this site are synthetic text produced by large language models. They are not verbatim statements made by any of the individuals analysed in this project — including Andrew Tate, Jordan Peterson, Esther Perel, Andrew Huberman, Jocko Willink, Brené Brown, or any other speaker in the corpus.
The generation pipeline is designed to keep the content as faithful as possible to each speaker's original YouTube transcripts. The full methodology is described in this undefended bachelor's thesis.