How Much Should We Spend to Reduce A.I.'S Existential Risk? (2025) [pdf] (www.nber.org)

A recent working paper by Charles I. Jones from the National Bureau of Economic Research explores the pressing question of how much society should invest to mitigate the existential risks posed by advanced artificial intelligence (AI). Drawing parallels to the financial measures taken during the COVID-19 pandemic—where the U.S. effectively spent around 4% of its GDP to combat a mortality risk of 0.3%—the paper suggests that a similar investment could be warranted for AI, especially given that experts estimate the potential existential risk from AI at up to 10%. The analysis indicates that an annual investment of at least 1% of GDP is justifiable, with simulations suggesting that investments could exceed 8% depending on effectiveness and risk assessments. This research is significant for the AI/ML community as it frames the existential risks associated with AI development in concrete economic terms, advocating for proactive policy adjustments in funding. By highlighting the potentially catastrophic implications of unchecked AI advancement and drawing from historical analogies, the paper underscores the urgency for decision-makers to prioritize investments in safety and risk mitigation strategies. This approach not only calls for a reevaluation of current funding but also sets the stage for broader discussions about balancing the benefits of transformative technologies against their inherent dangers.