Interesting Development in AI...Except for power in an emergency...

Much better than this guy...

Better? I'd say Otto is doing much better. If you're familiar with the movie you know why Otto is smiling.

What's the name of the movie? Smile, You're On Otto Pilot?

Humans learn from many forms of experience; this learning is continuous/ongoing. Rules exist and are modified based on social values.

Robots learn from programs, but this learning is not continuous/ongoing unless the program sets-up ways for this to occur. Where do the rules enter in?

When "first do no harm" meets situations for which all options cause or allow harm, what does the decision tree look like?

Recent experience with robot-controlled taxis in San Francisco shows emergency situations that are generally avoided by human drivers.

I prefer not to have AI or robots in use.

--JMM

I prefer not to have AI or robots in use.

Me either. At first it will be a weapon for some carbon-based creatures to use against other carbon-based creatures. When AI becomes sentient, it may logically conclude these carbon-based creatures are too dangerous to keep around.

Can't say as I would blame them...

I think the major problem with AI is that either humans will abdicate their lives to it, or it will usurp our lives from us. What is life anyway? How can AI accommodate a conscious entity, if a lot of people don't even believe that consciousness exists?

This could lead to a runaway train (so to speak metaphorically).