Although long-term prediction of chaotic dynamical systems is a fool's errand, I'd be interested to see what running an AI designed to optimally fit a model to historic climate data and including all of the following as inputs: CO2 and CH4 levels, solar modulation of cosmic-ray influenced cloud formation, known oceanic climate cycles, deforestation, soot output from human activities, and aggregate urban heat island effects from human development, would come up with in terms of relative importance, and what it would give in terms of short term predictions.
The last three are interesting because they are anthropogenic, but except for the first (and that only via it's effect on CO2 levels) completely ignored in the alarmist climate models which assume greenhouse gasses uber alles as a starting point.
