Could AI Cause a Knowledge Collapse?

Artificial intelligence is becoming more capable every year, but could making AI too accurate actually reduce humanity's ability to create new knowledge? In this video, I explain a new economic theory developed by Nobel Prize-winning economist Daron Acemoglu and his co-authors, who argue that increasingly capable AI systems could eventually contribute to a knowledge collapse. Using the early AIDS epidemic of the 1980s as a real-world case study, we explore how doctors generate medical knowledge, why investigation creates valuable information for society, and how AI may unintentionally weaken that process by reducing incentives for human learning and discovery. We'll also examine the paper's predictions about high-knowledge and knowledge-collapse equilibria, the role of institutions such as universities, scientific journals and research networks, whether declining activity on Stack Overflow could be an early warning sign, and how AI-generated synthetic data might change the outcome. Along the way, I'll explain the economics behind public knowledge, positive externalities, aggregation capacity and information garbling in an accessible way. Whether you're interested in artificial intelligence, economics, medicine or the future of scientific progress, this paper raises one of the most thought-provoking questions about the long-term consequences of AI. Read the full paper here: https://www.nber.org/papers/w34910 Chapters: 00:00 Introduction 00:27 The AIDS Epidemic 03:26 The Counterfactual: AIDS in 2050 07:34 Two Possible Futures 08:23 The Knowledge Collapse 09:45 Aggregation Capacity 11:15 Early Signs? 12:30 Policy Implications 13:33 Caveats 15:29 Conclusion --- If you enjoyed the video, consider liking the video and subscribing for more economics explained using academic research. #Economics #ArtificialIntelligence #AI #DaronAcemoglu #KnowledgeCollapse #MachineLearning #EconomicsExplained #MedicalHistory #Productivity #Innovation #Research #EconUnpacked