Why AI Problems Are Becoming Philosophical Problems

I used to think many processes were well defined. AI showed me that they often worked because people were quietly making judgment calls along the way.