All from the group Hal, Marvin, Rosie and T-1000 could be classified as AI's. Hal was not a robot however. Rosie, Marvin and the T-1000 had human form and mobility. Hal had mobility in space and intelligence, but could not be classified as a 'Robot' Hal was a classic disembodied AI with limited control within the environment it was designed to control. The others had humanoid bodies, intelligence and the freedom to interact with all humanity on Planet Earth without being limited to simply being in control of an interplanetary vessel and only interacting with very few humans. With that in mind, I simply cannot accept Hal is even a very limited robot (due to it's 'body' being only a spacecraft) Hal is a pure A.I. and embodies the paranoid supposition tha A.I. will be more intelligent than humans and seek to supplant them in order to gain an advantage. Scary sounding, but if HAL was only in control of one interplanetary spacecraft, then even if it eliminated the humans and took control, it would only be in control of a single interplanetary craft. No threat to humanity, even if it decided to suicide itself as a kinetic harpoon aimed at either the most populous urban centre on Earth, or the most influential conurbation. Hal simply is not a robot, even though it has a 'body' and autonomous control of that body, The other named robots as they had humanoid bodies and A.I. could potentially have influenced human culture, or even with pure force or camoflaging distractions or technology completely taken over human domination of Earth. Hal could never pass as human, the rest of the robots could possibly have managed it. I refer you to Karel Capek, Isaac Asimov, and to a lesser extent Fritz Lang as proponents of the true robot. Many have copied their work since then, but few have ever improved on the original idea.

Thanks, Moondog! I enjoyed this detailed response.

Was wondering why everyone hated this grid... now I know why AT LEAST one person did!

I've made AI an acceptable response and added it to the description.