The Theory of Human Agency (Is Agency Emerging in AI?)

Understanding AI Agency: Is Agency Emerging in AI?Mike Monday & Jon Gillick dive deep into the philosophical and practical implications of AI agency and reasoning. We explore why human psychology often misconstrues AI capabilities and the importance of precise language in AI research.Mike discusses his journey in understanding AI through listening to the Machine Learning Street Talk podcast, emphasizing the need to accurately define terms like 'agency' to avoid overhyped fears or evangelical expectations.Mike Monday then reveals his new Theory of Human Agency, a functional definition of agency that is testable and verifiable.What separates human intelligence from AI?00:00 The Emotional Experience with LLMs00:31 Diving into AI and Machine Learning03:09 Understanding Reasoning in LLMs05:44 Defining Agency in AI12:44 The Debate on AI Agency and Safety22:46 The Theory of Human Agency:23:22 Component 1: The Will to Survive24:23 Component 2: Perception24:44 Component 3: The Will to Act26:02 Component 4: Communication27:10 Component 5: The Will to Become28:03 Component 6: Persistent Memory29:53 What Does "Will" Mean?36:49 The Meat vs Silicon Fallacy38:19 The Future of AI and Human Agency47:54 Conclusion and Further ResourcesRecursive Intelligence (Mike's Substack): https://mikemonday.substack.com/Machine Learning Street Talk "How Do AI Models Actually Think?": https://youtu.be/14DXtvRJeNU?si=GcHAgBnrak21DTVC Get full access to The Monday Stack at mikemonday.substack.com/subscribe