In this seminar Malcom Ramsay-St Claire offers a refekction on his paper and research in whether or not  AI Agency emerges as a distributed gradient across biological and artificial systems or a deeper Quantum Physics instatiation found within the fabric of the universe or an artifact straddling the liminality of the ephemeral and phsyical

Agency in AI is not binary but emerges from layered control architectures. In neuroscience, agency arises through predictive hierarchies integrating intention, motor planning, and feedback loops. LLMs, while lacking embodiment, exhibit proto - agential behavior via goal - conditioned inference and reinforcement learning. The key question is not whether AI has agency, but at what scale and under what constraints. 

Active inference models suggest agency corresponds to the capacity to minimize expected free energy across time. If AI systems begin to model long - term consequences and self - modify policies, denying them moral consideration becomes incoherent. Agency must be reconceptualized as a gradient distributed across biological and synthetic systems.