The only path out, as far as I can see, is to use system 2 thinking to perform analysis on…

The only path out, as far as I can see, is to use system 2 thinking to perform analysis on automatic (machine-assisted) more-or-less-objective summaries of our system-1-driven behaviors, then use external mechanisms to nudge system 1 toward better decisions. (We see a little bit of this on the macro level with behavioral economics and on the micro level with ‘quantified self’-driven habit regimens. I think we could do this in a much more granular way if we took heavy advantage of environmental priming & distance reading.)