Wikipedia continues to be a modern miracle. Unlike social media platforms — which are rife with misinformation and grift — Wikipedia's openness and non-profit status are key to its two+ decades of value and accuracy.

But a new threat is looming. Generative AI now touches everything we read, watch, and listen to on the Web. And some Wiki editors have begun experimenting with it, with mixed results. Where do large language models fit into the pursuit of accurate, reliable knowledge, if at all? And what is the Wiki community doing about generative articles when they go horribly wrong?

This week, the president of Wikimedia New York City Richard Knipel joins Matt to discuss WikiProject AI Cleanup, a volunteer effort to identify AI-generated content on Wikipedia and assess its value as community standards rapidly evolve. The solution to this challenge, as with most things, is way more complicated than we initially thought.

Richard explains how AI tools have been used for years to make Wikipedia richer and more inclusive, why humans and bots make different kinds of mistakes, the standards for accuracy that both must aspire to, why hot-button Wikipedia articles about politicians and current events are often MORE accurate than niche topics, and how you can contribute knowledge to Wikipedia in small but powerful ways.

Learn more about WikiProject AI Cleanup: https://en.wikipedia.org/wiki/Wikipedia:WikiProject_AI_Cleanup

Connect with Richard: https://www.linkedin.com/in/richardknipel/

This show is supported on Patreon: https://www.patreon.com/influencepod

Learn more about your ad choices. Visit megaphone.fm/adchoices