Most of you guys have probably seen Discovery Channel's "Alien Planet," depicting the adventures of a couple of futuristic robotic probes exploring an extra-solar planet.

At one point, one of the probes rears back as if startled when it registers an unexpected, nearby movement. It seems kind of comical on first blush, but of course it'd actually be good practical programming since living creatures exhibit the same behaviour for good reason (the desire to keep on living, I suppose).

My personal experience with AI is essentially non-existent, but in first-year comp-sci I had to write a program to solve a maze, and allow the user to watch the solution real-time (good old Pascal). I threw in some pseudo-random decision making because I didn't want the program to solve the same maze in the same way, time after time, and found that the apparent behaviour of my virtual mouse (Algernon, of course) started to became eerily lifelike...and with shockingly little (though not recursive) code.

Maybe I was watching him with biased eyes. I was his daddy after all.