Ethics in robotics: interview with Dr. Michael Anderson

In this talk

1. 0:12

   Why preserve human minds

   Anderson argues that Earth and life here are impermanent because the Sun will eventually expand and consume life. He wonders whether the patterns of electrical activity in human brains could be captured in a more durable vessel as a backup.

2. 1:38

   Robots as a research tool

   Anderson says he initially expected his work on AI ethics to be about software, but found existing robots and tools made practical experiments possible. He describes using a robot as a research tool to explore whether people’s control of it could be improved through bioethics.

3. 2:55

   Learning duties from cases

   Anderson proposes using cases where ethicists agree on the right action to learn the underlying duties and principles. For example, a robot might need to charge while a patient needs medication; giving medication may take priority because it prevents harm.

4. 5:12

   Applying principles to robot actions

   A learned principle can rank possible actions, with equally preferable actions sometimes remaining. Anderson says the approach is manageable for robots with limited goals, such as elder care or store inventory, because they have fewer actions and perceptions to consider.

5. 6:39

   Put ethics before deployment

   Anderson’s guidance is to keep robots out of situations where the ethical questions are not yet clear. He says ethical understanding should come before placing robots in those settings.