So with software that makes decisions, could we start making them accountable before AGI? Without AGI there's no possible preventative measure, is there? Our M

So with software that makes decisions, could we start making them accountable before AGI? Without AGI there's no possible preventative measure, is there? Our ML systems could work like a carbon tax, where it has to incorporate costs/fines/etc for bad decisions so it makes similar decisions less frequently.