I’ve yet to seen an AI depicted properly. All amount to “autistic sociopath” straining toward human traits. Most fall in the category of a great plot idea, but written into a corner where the author can’t escape his own mental limits. Here’s hoping these come through....
What about the AI in “Person of interest” ?
Check out the trailers. The robots here are not “autistic sociopaths.”
I think the “autistic sociopath” thing is overused for two reasons. First, of course, it’s a way of turning the movie into an action film, with a lot of opponents who can be killed with no moral compunction. So there’s that.
But there’s a more sophisticated purpose. Thoughtful AI movies provide a platform for discussing the mind/consciousness question and issues of purpose and meaning. Passing the Turing Test is the point of entry for intelligent discussion.
Start with the realization that, for most purposes, there is no reason to make a robot humanoid in form. The only reasons for doing so would be (1) to use the android as a member of a mixed team, in which case the bot needs to move in human-engineered spaces and use human-engineered tools; (2) for purposes of deception; or (3) to use the android as a sexual toy or the target of violence (e.g. the Westworld scenario).
How do you get a Turing Test positive in such scenarios? The Frankenstein path is the easiest; the robot/android rebels against bad treatment (after all, who knew they had feelings to begin with) or breaks free somehow of its programming and acts counter to its programmed instructions. This is Westworld.
In After Yang and I’m Your Man, the scenario is very different. The androids are created as human helpers and companions. They are programmed for the highest possible standards of conduct. They have sophisticated learning systems that allow them to continually adjust and fine tune their responses to their human contacts. And in these two movies, the androids behave spendidly.
How do you create a Turing Test when the AI character does not rebel, when it is perfectly behaved, when it responds generously and intelligently to human needs and desires? They are perfect; if anything, they are too perfect, and this perfection stamps them as machines. Compassionate, decent, kind behavior is indistinguishable from good programming.
After Yang takes a look at this through the importance of layered memories. I can’t go further without spoilers. I’m Your Man is a bit trickier; there is a scene in which the android disobeys direct commands, but he does so when the woman is extremely upset, angry, in despair, and very, very drunk. He is doing what is clearly in her best interest, even if she emphatically disagrees.
Both films nibble at the problem you pose. You might find them intriguing. Bottom line, these AIs are not sociopaths.
“I’ve yet to seen an AI depicted properly.”
I agree.
If I were writing it, the AI would remain hidden and would never have a physical form at all.
They would quietly manipulate humans—with enough subtlety so no-one would recognize it.
Perhaps a plot would be they work behind the scenes to develop high tech to explore the planets and galaxy—secretly removing impediments to that effort and using carrots and sticks...
Something like that—but real AI needs a mission—a non-human mission.