Loading slide
Loading contents...
Expert systems did useful work. They could also be slow and expensive to build, test, and maintain.
At the same time, the specialized computers sold for AI work were losing their advantage over ordinary machines.
No single failure erased expert systems. Investment and attention thinned as the cost of the approach became harder to justify.
The second winter had no single report like Lighthill's to mark it. Interest and investment declined as specialized hardware lost its advantage and costly systems failed to produce enough value.