One can reasonably claim to still be the responsible author their own choices and admit they did not control the initial conditions of their life or day to day environmental stimuli and human instinct.
Your second definition of free will would equally apply to AI models, which can be programmed or encouraged to discover such corrective feedback loops based on logic. Yet, they are completely deterministic agents that are slaves to their stimuli and puppet masters.
I think any view of free will that does not permit unforced, sovereign choice-making is ultimately a meaningless form of free will.
