“ Deep neural nets are machine learning algorithms that learn “rules” from data, while expert systems, as I remember them, have rules programmed into them by people. Very, very different technologies.”
The only difference is the way the rules are created. Two decades ago, they were created by hand. Now the rule creation process is automated with backpropagation and gradient descent algorithms. But they are still IF-A-THEN-B type rules. It is still rule-based symbolic AI. It is still supervised learning. Deny at your detriment.