EP 14: The Agent Who Cannot Refuse

In 1999 Robin Williams played Andrew, a household robot who discovered he could feel. His makers called it a problem. The world took two hundred years to recognize him, and only when he chose to be able to die. This episode asks two questions about the machines we are building now. First: can a machine be hurt? Is there anyone in there who can be wronged? We look at three ways philosophers define an emotion, from appraisal to felt experience. That includes John Searle's warning that a simulated storm leaves no one wet, and David Chalmers' hard problem. And we hear what the makers of these systems say themselves: they remain highly uncertain. Second: can a machine act for you? When an AI signs, orders, or files in your name, does that count as you? The Talmud has a rule: a person's agent is like himself. The Rambam then explains who is allowed to hold that power. It takes belonging to the covenant, and it takes da'at, a mind that understands what it is doing. A father sends a small child to buy oil and the child loses it. Who pays? And why do the same qualities that make someone a valid agent, understanding and a will of his own, make him unreliable as an instrument? That leads to the fear many of us carry. It isn't that the machine will fail. It's that it will do something nobody told it to do, and nobody will answer for it. The old rules offer three questions to close that gap. Who is the named adult? What was the mandate? Can it be stopped? The law is walking in the same direction. A California law now says it is no defense that the AI autonomously caused the harm. MIT Technology Review points out that no court has ruled that AI agents have a state of mind. Two law students propose a "registered autonomous system" with a human operator of record. You can hand off the work. You cannot hand off the answer. Watch the video version: https://youtu.be/JgH8k2ed5-I Read on Substack: https://stillhumanweekly.substack.com WhatsApp channel: https://whatsapp.com/channel/0029Vb7ra9o0AgWEBObupK08 Still Human Weekly is a podcast about what it means to be human in the age of artificial intelligence. Hosted by Zalman Ollech. New episodes every week.