The Returns to Science in the Presence of Technological Risk

Scientific and technological progress has historically been very beneficial to humanity but this does not always need to be true. Going forward, science may enable bad actors to cause genetically engineered pandemics that are more frequent and deadly than prior pandemics. I develop a quantitative economic model to assess the social returns to science, taking into account benefits to health and income, and forecast damages from new biological capabilities enabled by science. I set parameters for this model based on historical trends and forecasts from a large forecasting tournament of domain experts and superforecasters, which included forecasts about genetically engineered pandemic events. The results depend on the forecast likelihood that new scientific capabilities might lead to the end of our advanced civilization - there is substantial disagreement about this probability from participants in the forecasting tournament I use. If I set aside this remote possibility, I find the expected future social returns to science are strongly positive. Otherwise, the desirability of accelerating science depends on the value placed on the long-run future, in addition to which set of (quite different) forecasts of extinction risk are preferred. I also explore the sensitivity of these conclusions to a range of alternative assumptions.