I was surprised by David Laibson's belief that the majority of policymaking in the future will be done through machine learning instead of the traditional 20th century scientific method. I am surprised by his belief considering machine learning skips straight from data to prediction, unlike the scientific method which includes the rule, theory and explanation. I would've thought that the lack of the rule, theory and explanation of machine learning would make it largely unpalatable to the constituents of policymakers, limiting the use of machine learning in policymaking. However, machine learning has become increasingly a part of our technology, and people are increasingly using it. So, perhaps my feelings that people would not be accepting of this new technology at our highest levels of government are unreasonable.
Link to video: 1. Behavioral Economics - LabXchange