Online conference by Jan Eißfeldt, External Applied Complexity Fellow at the Santa Fe Institute, and fellow for a one-month writing residency in May 2026 at the Paris IAS as part of the "Paris IAS Ideas" series.

The "Paris IAS Ideas" online lecture series offers short and stimulating presentations by researchers from the Paris Institute for Advanced Studies, marking the start of their month-long writing residency.

Session exclusively online and in English.
Free upon registration.
Registration is required via the form at the bottom of the page to receive the login link.

Presentation

Being both the cornerstone data set used to build AI models from Boston to Beijing and a complex collective intelligence project, Reddit plays an outsized role in shaping the development possibilities of emerging AI approaches and in turn is uniquely impacted by them. The collaborative platform, curated by mostly self-governing volunteer communities in many languages that partner with the platform provider, the Reddit Inc., is employing adaptive coordination strategies to navigate the impact of AI model outputs flooding digital content ecosystems.

In the digital economy, model outputs are competing for monetizable attention with academia and journalism, bringing down the market value of traditional high quality content production. Historically relying on the latter, Reddit's epistemic ontology of voting on sources is guiding the project's collective peer production of its content. This CI-produced content is itself the most widely used high quality data set to train AI models. The interdependent dynamic has created an epistemic feedback loop between collective and artificial intelligence with distinct enabling conditions and impacts across languages.