Edward Shortliffe and Colleagues Develop MYCIN at Stanford

As mentioned, it's the early 1970s and expert systems are all the rage. Learning has fallen out of favor temporarily in the West.

If you wanted AI funding at this time in places like the United States, it had better be symbolic AI.

Work on the symbolic MYCIN began at Stanford in the spring of 1972. It was a collaboration of medical student Edward Shortliffe, computer scientist Bruce Buchanan and clinical pharmacologist Stanley Cohen that became Shortliffe’s 1974 dissertation.

The expert system program advised physicians on antibiotics for bacterial blood infections and, later, meningitis. Its knowledge sat in if-then rules drawn from Cohen and infectious disease physician Stanton Axline, 200 at first and 450 later. MYCIN questioned the physician, reasoned backward from possible conclusions and could show which rule prompted each question. In a blinded 1979 test on ten meningitis cases, eight outside specialists rated it above all nine human prescribers. However, the system was never tested in an actual hospital setting on real people.

Even symbolic AI had limited trust capital, especially when it came to people's health.