I came from Czechoslovakia to the Stanford AI Lab in October 1967. This lab was one of the three AI labs in the USA and was under the leadership of Professor John McCarthy.
The basic philosophy of Professor McCarthy was that AI was about representation of knowledge and that this representation was symbolic. The language we used for the representation was LISP. To the credit of Professor McCarthy, he recognized that perception and robotic interaction with the environment was equally important as reasoning strictly on symbolic information. Hence we faced the problem of how to systematically convert the measurements or observations into symbols. What is an edge, straight line, circle, a cube, and so on? This is still an open problem.
The tradition that was set at that time (and it has prevailed) is the foundation of a good engineering science: every good theory needs experimental verification. As we go on and understand more complex phenomena, the experiments reflect this complexity.
I implemented this tradition in the GRASP laboratory during my 30 years at the University of Pennsylvania in Philadelphia. Furthermore, coming from a background of Control Engineering, we recognized the need in building intelligent systems the importance of controlling the data acquisition and introduced an new paradigm: active perception. We stated that we just not see but we look and we not only touch but we feel.
AI has come a long way. The students of AI are sophisticated in both discrete and continuous mathematics, including a recognition of the role of uncertainty. This is necessary because of the increased complexity of problems that we need to attack.