Expert Systems

Expert Systems capture and emulate human decision making via logical rules. They are basically hard coded AI with features to make them easier to update via non-programmers. E.g. rules can be expressed in simplified format and stored in a database. This is easier than actually coding a bunch of IF-THEN statements and automatic organization, searching, and application of the rule statements can be used.

More advanced systems applied "Fuzzy Logic" to express rules not as black or white, but in terms of confidence levels, with answers being delivered in order of the aggregate support for that answer provided by input to date.

The biggest disadvantage is the need for experts who actually understand the rules they are applying to make decisions, and then correctly enter those rules into the system in a way it could apply.

Common programming languages used to implement expert systems are LISP and Prolog.

See also:

• http://www.cs.utexas.edu/users/novak/tmycinb.html A tiny version of the famous MYCIN^ expert system written in Lisp. Included data helps to identify rocks and snakes.
• http://www.primaryobjects.com/2018/07/23/logical-based-artificial-intelligence-and-expert-systems/ A logic based expert system library written in JavaScript.