As a good starting point, I asked ChatGPT the following:
Develop a theory of free-will based on first principles of Active Inference. Provide references.
I found the answer too vague, and it doesn’t really explain anything:
“providing a nuanced understanding of free will that integrates both unconscious predictions and conscious control.”
Ok, but if conscious control is constrained by prior probability matrices, albeit of a higher level, is it really free?
In other words, it needs a lot of work.
