Most environmental issues are fundamentally policy issues that currently are poorly tackled because of socio-economical and political reasons. To me it's unclear whether ML can (or should) fine-tune policies that struggle to even be brought up to legislators / the public.
If so, we should use ML to make a _better_ case for tackling such issues. Maybe we should chat with the orgs involved in lobbying policymakers and provide data analysis tools to strengthen their cases. Would be great to have a joint workshop on this at NIPS/ICML/AAAI/IJCAI.
Well said & agree. Would love to participate in a new virtual conference (airplane travel wrong symbol) focused on the topic of Flora-Fauna-Human-Computer-Interaction (FFHCI) - Bringing together CS,ML,HCI people with NGOs, activists & policy makers to discuss climate change challenges & solutions.