I would argue no free will and no conscious part, human brain is just a highly advanced chemical calculation unit that ran on basically non-stop random generated tasks, just like the hardware counterpart we use on making AI, it would be hard to tell if AI, specially with the latest model that can store new knowledge long term and when continuous random prompts generating power are given to, don't actually have the illusion of/ covering their free will. Albeit they now still thinking the world around them very slowly and might unable to be understanding by human, they might still hold some level of self preservation and gain (randomly generated) motively reactional ability like cats and dogs. That is the dangerous part, human never able to tell when if it is already too late until its been proven too late...
@Saviliana