One of the hardest things for any algorithmic leader is knowing when do nothing at all. This is not an entirely new dilemma. Test pilots in the early days of the space program, struggled with the idea of not having manual controls - even when their own interventions led to deadly mistakes. So just when do humans make good decisions? To get to the bottom of that, I chatted with Jason Collins, a behavioral economist, who has written extensively on these ideas at the Behavioral Scientist, and currently runs the data science team at a major financial regulator. He previously co-led PwC Australia's behavioral economics practice.
