Experimental longtermism: theory needs data

Abstract

The article proposes experimental longtermism, arguing that the field of existential risk reduction lacks sufficiently robust feedback loops to improve its models and predictions. It contends that the COVID-19 pandemic offered a rare opportunity to test these models in real-world scenarios, by applying them to the pandemic’s challenges. The author and their collaborators intervened to mitigate the impact of COVID-19, viewing their efforts as experiments to assess the validity and effectiveness of their models. While direct impacts were observed, the primary value derived from these interventions was the acquisition of information, improving decision-making processes. The author suggests that more longtermists should engage in similar experimental activities to enhance the field’s understanding of real-world dynamics, ultimately leading to more effective and informed approaches to mitigating existential risks. – AI-generated abstract.