Summary of State of AI Systems in Practice  

* Speech synthesis, recognition and understanding

    * very useful for limited vocabulary applications

    * unconstrained speech understanding is still too hard

* Computer vision

    * works for constrained problems (hand-written zip-codes)

    * understanding real-world, natural scenes is still too hard

* Learning

    * adaptive systems are used in many applications: have their limits

* Planning and Reasoning

    * only works for constrained problems: e.g., chess

    * real-world is too complex for general systems

* Overall:

    * many components of intelligent systems are “doable”

    * there are many interesting research problems remaining