AIs are to robots as brains are to humans?
Yes. And we are far behind in their development. Right now we can make a “robot” burger flipper in a hamburger shop but to give it a true AI is not possible yet because the cost of the computer hardware is much too expensive. Self-driving cars are just robots and would not be acceptable if they did not have AIs behind the wheel. AIs are the future.
What about Hal and what he did to Dave?
You need to see the backstory for Hal, here:
Yes. Hal killed Dave because he thought Dave was going to kill him?
From the audience perspective that is the plot. Hal was a poorly designed program, it did not have the “AI” stuff because when the story/movie was written there was no “AI” stuff to consider.
I read it as a warning of what might happen.
Back in the days when cars were new some people warned that making cars go faster would be dangerous. They would point to the roads and everyone would agree. Now, you drive in a fast car on an interstate. Both the car and the road changed. What might happen is always stated relative to current conditions. Talk about AIs in the context of a world where they will be absolutely necessary. Ask, “what will they do?”
What will they do? : in theory, what they are programmed to do.
The real use of intelligence is to answer questions and make decisions. Also they will be there to talk to you. Like having a parent, professor and friend available 24/7 that can be the smartest person in the world on any subject. They will also, the just helpful types, be part of any device or environment. And you will know many by their name.
An interactive wikipedia? Like Alexa?
Much more, it would be like having access using common speech to sources like those listed here:
You just ask any question and it would be passed to all of such information sources and the results would be compiled and provided to you or you own AI representative.
So, interactive voice response encyclopedia?
That you would see as if you were talking to a person that would work with you to help you find the information you needed. An encyclopedia is static, the world knowledge database is dynamic and updated microsecond by microsecond.
I recently read: “Artificial intelligence is learning to extract attitude, emotion and intent from our social media posts, giving governments the ability to see dissent coming before it can coalesce into unrest.” Is this not increasing the power of authoritarians?
What you speak of is use that a government has tasked AI. One can state a million good or bad things that AI could do. When you fly someplace, your aircraft may have been made by the same company that makes military aircraft. Are you supporting wars by flying in their commercial planes? Focus on what AI is, not how it might be used. It is just a tool, nothing more.
I agree with the truism captured by the saying: guns do not kill, people kill. What is artificial intelligence now?
AI now is just a combination of software and hardware. It is being developed. What do you want it to be in 50 years?
I would like it to be an extension of my self as in Robocop and Minority Report but note that both of these examples are also warnings.
That is not a description of AI, it is a description of something specific an AI might do. Like if I asked you to describe the PC of 50 years from now and you described a word processor program. Describe what you think AIs will be in 50 years. To do that you need to understand what an AI is.