In this episode, Ross Dawson and Jason Burton explore the transformative role of large language models (LLMs) like ChatGPT in enhancing and reshaping collective intelligence. Jason shares insights from his recent paper, co-authored with 27 other experts, which examines how LLMs act as new aggregation tools, merging collective knowledge from vast online sources into coherent responses. The conversation touches on how LLMs can simulate crowdsourcing, provide fast and diverse answers, and even serve as facilitators in complex deliberative processes. This evolution raises questions about the benefits of such technology in fostering productive dialogues and generating collective insights. The discussion also addresses potential downsides, such as the risk of homogenized thinking when many users rely on the same LLMs, leading to diminished diversity of thought. Jason shares his concerns over the current landscape dominated by a few major tech players, suggesting that more diversified models could help mitigate these effects. The conversation concludes with a vision for future applications of AI in civic technology, where language models could bridge public discourse and bureaucratic decision-making, ultimately contributing to a more informed and engaged society.