I feel like this is the next Big Thing that movies will be about. Everywhere we look, we will be seeing either overt or metaphorical grappling with the coming threat of artificial intelligence.
The coming threat ... and opportunity?
You see, once you start seeing a lot of a certain something at the movies, someone needs to come along and upend that so it doesn't get stale.
And so before we get too much further into this piece, I better issue a SPOILER ALERT for both Timur Bekmambetov's Mercy and Andrew Stanton's In the Blink of an Eye.
Like, pretty big spoilers, so if you don't want to have anything more ruined than the subject of this post, and the fact that these two movies require a spoiler alert related to that subject, don't read on.
So the first of the two movies I watched was Stanton's latest, which I think came with some bad word of mouth. It's got a respectable 6.1 rating on IMDB, so I can't say I remember for sure where the bad word of mouth came from. Let's just say that any time you divide a story between three timelines, and one of them features cavepeople, you're in trouble.
Yes this gave me Cloud Atlas vibes a bit, but hey, I liked Cloud Atlas.
The pertinent storyline is not among the cavepeople -- no, this isn't some alternate universe where they imagine that AI existed among cavepeople -- but about a ship hurtling toward a distant planet in the distant future, which it will reach in something like a hundred years. The only human on board is the character played by Kate McKinnon. You'd think this is going to be one of those situations where Kate has to die and teach her offspring to run the ship, because it will take so many generations to get there that she will be long since dead by the time it arrives. But no, Kate has had her aging slowed down so that she's basically immortal -- a thing that became possible on Earth before we destroyed it, requiring a new home for the human race.
Well, not much of the human race. Actually only just her and a bunch of fetuses that are in suspended animation, waiting to be awoken at strategic intervals so there will be plenty of young humans still around when they get to the new planet.
She does have a companion, though: the ship. Or an AI on the ship. They may be one in the same.
A different sort of movie would take this in a different direction, and the AI would be some sort of HAL 9000 with sinister designs on this paltry quantity of remaining humans. Not this AI, though.
There comes a point on the journey when the plants are damaged and they are not producing sufficient oxygen to get the human cargo to its destination. There's a way to produce more oxygen, but it involves stealing so many resources from the AI that it will die.
And this AI, in a pleasant female voice, puts up its hand and volunteers for the suicide mission.
Kate has developed such a bond with this AI that it's a real death, and almost certainly presages the inevitable failure of the whole mission. I don't need to tell you if that's the outcome or not to make my point.
The point is, this AI does not, for a second, consider its own self-preservation in this scenario, even though it is quite clearly sophisticated and capable of what we would call emotion. In fact, that's clearest in its final request, having prepared for its unplugging in a way that is matter-of-fact yet also has an air of human deliberation to it. Before it goes, it says "Take care of the children for me," and you just want to cry. (I didn't, but it was an effective moment.)
Imagine my surprise when only nine days later I saw another movie about an AI in which the AI is actually the good guy.
For sure, the AI judge played by Rebecca Ferguson in Mercy is set up as an adversary of the character played by Chris Pratt, a detective strapped in a chair, where he must plead in his innocence in the murder of his wife or be executed in 90 minutes. It's quite the dystopian setup, but the evidence seems to suggest that the AI makes an effective judge. In this version of society, so much is surveiled, so much evidence is captured on film, and so much of this is instantly accessible to the AI that compiles it all, that basically this system is never wrong in determining the guilt of the suspect. It's in the same territory as the precogs in Steven Spielberg's Minority Report, which is the movie that this will obviously make everyone think of first and foremost.
As cold and detached as Ferguson's very good performance is -- I actually liked Pratt in this too -- it's clear that she's not a malevolent AI. In fact, it's clear that as the movie goes, she learns that maybe she can be wrong, and even seems to break protocol to help the man on trial -- which you'd think an AI would never do. In fact, you might say that without the AI judge's help, this man would have certainly gone to an early grave.
When artificial intelligence learns how to be "more human," we never hear about them becoming empathetic. We only learn about them developing megalomania or an outsized survival instinct. In most visions of an AI future, these are the traits they acquire from humans, not their capacity for good.
In the space of ten days, I saw two films that imagined a more three-dimensional AI than the many boogeymen who have come before them.
Having said that, the third AI movie I've watched this year -- in which the AI is not very good -- is Good Luck, Have Fun, Don't Die, which is easily my favorite of the three. So maybe our fears of AI are still better developed than our potential nascent trust of them.
No comments:
Post a Comment