I have a fantasy about June 26, 2001. I have a fantasy about a certain person, a die-hard, unapologetic Kubrick acolyte, who has come to witness the debut of Steven Spielberg’s A.I. Artificial Intelligence. There she/he sits, in the very first row of the very first screening…but not to watch Spielberg pay homage to friend and mentor Stanley Kubrick, who developed and largely fleshed out the original idea for A.I. (with a significant contribution from Ian Watson) before passing it on to Spielberg in the belief that the director of Close Encounters of the Third Kind and E.T. the Extra-Terrestrial could better navigate the film’s emotional beats. No, this person has come with an expectation, born of a certain over-simplified preconception of Kubrick, of Spielberg.
This person has come to witness his/her worst nightmare come true.
For the first two hours and change, it must have been hard. Kubrick and Watson are credited with the story, but Spielberg takes sole screenplay credit, his first in twenty-four years. As a result, A.I. is admittedly an odd mixture of styles. There’s a Spielbergian, fairy-tale patina to the story of David, a childlike robot played by Haley Joel Osment, who’s abandoned by his “mother” and subsequently goes on a quest to find Pinocchio’s benevolent Blue Angel, in the hopes of becoming a real boy. But an uncanny cynicism keeps creeping in around the edges, from the myriad, horrendous ways the passive, serene androids are dispatched in the garish Flesh Fair, all the way to David’s ostensibly Geppetto-esque creator (played by—who else?—William Hurt at his warmest and most paternalistic), who’s nonetheless unabashed about abusing his human-like androids if it will help him make a point, and whose ultimate goal is not a single, precious life-like boy, but an army of money-making affection-synthesizers rolling off the assembly line. My imaginary Kubrick supplicant must have sat there—fingernails digging into armrests, each crunch of popcorn withering to the flavor of pure ash in her/his mouth—waiting for that moment when their cinematic god would be betrayed, only to find him/herself thwarted at each beat.
And then, at precisely two hours, twelve minutes, and forty-three seconds, it came. And that person was finally able to leap up from her/his seat, and cry to the world, “I KNEW IT! I KNEW IT! YOU COULDN’T RESIST, COULD YOU? YOU HAD TO TAKE THE MASTER’S BRILLIANCE AND SLAP YOUR SPIELBERG SCHMALTZ ALL OVER IT. DAMN YOU TO HELL, YOU SONUVABITCH! DAMN YOU!”
Or something like that. Probably didn’t happen. No, definitely didn’t. I think we would have heard about it.
That aggrieved outburst, if ever it had come, would have been during A.I.’s coda, set two thousand years after the film’s main story, when a group of crystalline androids retrieve David from the bottom of New York’s Lower Bay and give him what he had been seeking from the submerged Blue Fairy for two millennia: One perfect day with a mother (Frances O’Connor) who finally tells him she loves him. It was there, in what a large chunk of A.I.’s audience perceived as those over-sentimentalized final minutes, that the film’s infamy was cemented.
For all the wrong reasons.
The standard rap goes like this: A.I. Artificial Intelligence should have ended a half-hour earlier, with David pinned underwater by a collapsed, Coney Island Ferris wheel, begging an effigy of the Blue Fairy to make him a real boy…begging ‘til the end of time. That was the grim, despairing image that—in the minds of many—would have paid appropriate tribute to Kubrick and his legacy. By this interpretation, the far-future coda was just Spielberg attempting to paint a happy face—with perhaps a single, poignant tear—on a much darker concept.
Oh, yes. So very Spielberg. But no.
For one thing, both Spielberg and Watson have confirmed that the coda was always part of Kubrick’s plan. Spielberg may have put his own spin on it, but, then, what of that? When one’s view is clouded by a reductive understanding of the famous director’s rep—Spielberg as the magical wizard of family-friendly film whose tales are spun of kitten dreams and candy floss—it’s way too easy to dismiss the finale as sappy melodrama.
But it’s no big news that Spielberg’s oeuvre, for all his weaknesses—which he does have—and strengths—of which there are many—was never just sentimental, soft, or sweet. His breakthrough film, Jaws, pushed mainstream cinema’s limits for explicit violence and gore; Schindler’s List was unsparing in its depiction of Nazi atrocities; the abduction sequence of Close Encounters is mounted as pure horror; hell, even sweet, dreamy E.T. begins with an unsettling sequence that put more than few eight-year-olds on edge. Spielberg has never shied away from the darkness, and when he’s tried, it hasn’t gone so well. (Is Hook anything other than Spielberg and Robin Williams collectively crying, “Who do I have to fuck to get out of this arrested-development, man-child horseshit?”)
Admittedly, A.I.’s final act has been reconsidered and reanalyzed over the years, with critics conceding that it’s not merely the gooey curtain-dropper it appears on the surface. But even there, the focus tends to settle specifically on what the end means to David’s journey, and his implied death after having achieved his mission to become a real boy. But there’s something even darker in that finale that seems often to be missed. Something that would resonate strongly with what we know about Kubrick, his views of humanity, and his thoughts about its ultimate fate.
There’s a moment at the end when one of the androids—which, parenthetically, bear a striking resemblance to the Giacometti sculptures that Kubrick at one point had planned to use as models for the never-seen aliens of 2001: A Space Odyssey—explains to David that robot society has been searching for any connection to the long-dead human race, culminating in futile attempts to recreate the species. The question is: Why? Why would they need to do that? There are no overt indications that the machines David interacts with are academics, or archaeologists—there’s no suggestion of a museum or zoo in which the once-dominant inhabitants of Earth would be enshrined. As much as all that background could be inferred, I don’t think it was intended.
Buy the Book
A Psalm for the Wild-Built
Think of the timeline here: We fade out on David begging the Blue Fairy to make him a real boy. At that point in the story, he exists in a world where machines exist to serve their human masters, passively, uncomplainingly, to the extent that they’ll willfully go off to their own, violent demises if so ordered. Even David, built to (at least exhibit) love, is nothing without a human—a mother, specifically—with whom to interact. We then fade in, returning to the world two thousand years in the future. The human race is now effectively extinct, but to the surprise (and, possibly, disappointment) of Terminator fans, the fatal event seems to be a self-inflicted, environmental apocalypse (remember those rising tides?), not a superior robot army grown tired of humanity’s foofaraw. In fact, as technologically advanced as these crystalline, graceful androids appear, their demeanor is not much different from the compliant machines of David’s time.
And it’s because, I think, they aren’t different. Humanity has died out, but without ever granting their electronic progeny full sentience. As much as a Singularity of a sort may have been attained—at least to the extent that the robots have enough autonomy to improve their own technology—the androids’ sole motivation remains that of being of service to humans. And so (to fall back on a recently-coined Rick and Morty-ism), an Asimov Cascade occurs: In the absence of humans to serve, the androids must recreate humans, so those humans may be served. What they find instead is David, “the enduring memory of the human race.” His synthesized humanity is just convincing enough that when the (presumably) head android hears David’s demand to have his mother recreated, and responds, “Give him what he wants,” it’s not out of empathy, but because something that at least exhibits the recognizable markers of humanity has at last delivered what the robots have been seeking for two millennia: an order.
Does David’s perfect day with his mother seem overly sentimental? Yes. Maybe that’s Spielberg falling back on old habits. Or maybe it’s Spielberg deliberately using those habits to posit what machines, with no emotional history to speak of, would synthesize for a robot whose own emotional history is just as superficial. When viewed that way, the final fade out, as the lights dim in the recreated home and David drifts off “to that place where dreams are born,” is not the bittersweet culmination of a young robot’s wishes fulfilled, but something far bleaker. David—the machine that would be a boy—has delivered the very last order his fellow robots will ever fulfill, and what we have witnessed in that “perfect” day is a simulation of humanity, the best these highly advanced robots can manage, but falling far short of the real thing, almost to the point of mockery. David’s departure then represents the closing chapter of a rare and precious commodity: the human spirit. The Earth is left in the custodianship of brilliant machines, totally lacking in motivation beyond their own propagation. Gone is inspiration, imagination, passion. Gone now, finally, totally, is true humanity, never to be recovered.
A.I. Artificial Intelligence is, to put it mildly, a problematic film. Stanley Kubrick wasn’t completely wrong in feeling that Steven Spielberg was a fitting inheritor of his tale, and Spielberg wasn’t completely wrong in the steps he took to reconcile the ideas of his mentor with his own vision. Still, despite Kubrick’s wit and Spielberg’s bent toward darkness, it’s an uneasy mix, Spielberg’s dreamy, fairy-tale mise en scene dancing precariously with Kubrick’s treatise on how humans could inadvertently engineer the demise of the thing that makes them human. Still, looking back at the film two decades later, there should not be any argument that Spielberg didn’t honor his commitment to Kubrick, right through to the very last seconds. In so doing, he gave us a compelling reminder of what we should cherish about our humanity, and the efforts we need to expend to make sure it doesn’t gutter out and die like the final dreams of a robot who wished to be a real boy.
I’m no good at being noble, but it doesn’t take much to see that the problems of a highly sophisticated child android don’t amount to a hill of beans in this crazy world. But your opinions do! I’ve given my take on A.I., now it’s time for you to give yours. Keep it friendly, keep it polite, and feel free to comment below.
Dan Persons has been knocking about the genre media beat for, oh, a good handful of years, now. He’s presently house critic for the radio show Hour of the Wolf on WBAI 99.5FM in New York, and previously was editor of Cinefantastique and Animefantastique, as well as producer of news updates for The Monster Channel. He is also founder of Anime Philadelphia, a program to encourage theatrical screenings of Japanese animation. And you should taste his One Alarm Chili! Wow!
I will always remember the “Foxtrot” comic strip in which the two boys fall off a cliff. As they are falling, one says “My life just flashed before my eyes, and ‘A. I.’ is still too long.”
Maybe people wouldn’t dislike the coda so much if the film had been shorter.
Spielberg sometimes has a problem with ending his movies. “Lincoln” being a prime example. It should end with Lincoln leaving for Ford’s Theater, saying “I don’t want to go, but I must.” However, we then see Lincoln’s son going into hysterics at a different theater when news comes of his father’s assassination. Then, we get a flashback to Lincoln making an uplifting speech. All three scenes would have been good endings, but not one right after the other.
As I vaguely recall, “A. I.” had something of that problem. We felt like the movie had ended, but no, it hadn’t. And no, it hadn’t again.
I’m close to your fantasy person in that I like Kubrick a lot and don’t like Spielberg very much… but my reaction was quite different. The *first* act was the nightmare—a bad Kubrick pastiche. The middle was pretty good, and the ending was gloriously insane.
There are some great scenes in AI. It’s almost a very good movie. Alas, it does feel like a fudge – even allowing for how Spielberg’s ending might be more ambiguous than it seems at first glance.* That’s partially due to the fact that it’s the work of two quite different directors, but maybe also because Kubrick didn’t know what kind of movie he was making either? One of the best things about Kubrick was that he always seemed to have a very clear idea as to what sort of movie he intended to make. He then went ahead and made it. Here? Not so much.
* and it’s hard to shrug off the impression that Spielberg just couldn’t resist some sort of happy ending, no matter how ambiguous.
@3 except that ending was in Kubrick’s draft.
Although I knew the script had been in development hell for years, I hadn’t realised until I read this article that Kubrick effectively handed the project over to Spielberg four years before he died – in effect, admitting that he didn’t know what to do with it. Maybe the ending was a factor? I think Kubrick overthought the movie*, found himself in a creative cul-de-sac and passed the baton onto somebody else.
* Pinocchio (his provisional title) and Supertoys Last All Summer Long are tonally very different stories. He was setting himself up for a fall, imo.
I read the final scene differently, based on the grown-up robot’s remark that David was made “too young.” He was made to exactly emulate a human child–and I think he really does feel fear, loneliness, and love–but only certain parts of a child’s personality that certain sorts of adults* like to interact with. He might be extremely bright in human terms, for all I know; given the right environment he might be the equivalent of a prodigy in math or ceramics or something. But emotionally he cannot leave the stage at which the parental figure is the axis of the universe. That is literally what he is for.
It’s possible that, given enough information, they might have been able to turn that off and reprogram him to be fixated on a robot who was actually there and willing to take on the responsibility. But they found him on an archeological dig. What are the chances that they will ever find enough information about how to help him?
So they make a fleshbot, program it with everything David remembers with happiness about his mother, and give him one last, happy day before they remotely hack his mind and give him a happy half-dream to savor as they induce him to shut down again.
In one day they are kinder to David than all of humanity was in his whole life. And, I like to think, they carefully preserve his body and his brain, in case they ever do find a way to help him get past what humankind did to him.
*Because there is a certain chain of decisions that goes into purchasing a programmable child that reacts just like a human and loves only you, but has no legal identity. David’s mother may have been operating from grief and incomplete information, but she did purchase a child.
@6 Purchase? In the film, the family receives David as a test of a prototype, and Monica is initially opposed to the idea.
Does.anyone remember a film called DARYL from the 80s about a boy android escapes from a secret lab and a couple adopt him as their son, and his eventual realisation that he isnt human,and so how to save him as a human rather than a failed military experiment?
It’s a fantastic film and so is AI,hell make it a double bill!
There’s so many profound particles of cinema in it. I feel it gets criticised too much. The last part was powerful to me. How we deal with loss, grief and an with a possible oedipal element. What we wouldnt do to have one perfect day with our passed away loved one.
@8
I remember DARYL. And if memory serves, Tor.com did a retrospective review of the movie a few years ago.
@8, 9–I had forgotten all about D.A.R.Y.L., but you’re right! The loving discussion of the movie can be found here :)
@7: I stand corrected. She is not in the group that would eagerly buy a David robot so they could have a child frozen at the age of adoration.
But she chooses to dump David by the side of the road, like an irresponsible dog owner getting rid of a troublesome pet, because she is able to believe (enough, and for a sufficient number of minutes) that he doesn’t actually feel–that he’s just some kind of mechanism that’s malfunctioning in an upsetting way. Which puts her in another group of people who should never have access to a David robot.
I think this is exactly backwards — the “Asimov Cascade” was the creation of less and less real and convincing decoys/dopplegangers/simulacra/eidola by each generation of decoys. But the post-post-Singularity robots had no problem creating a human simulacrum that was perfectly convincing. If all they really wanted was to have humans to order them around, they could have created exactly that; they would have already had that. So I’m pretty sure the robots have no such need.
I would have preferred that the future robots had offered to David the option of upgrading him out of his programmed codependency, but maybe that wouldn’t have been Kubrickian enough?
Love this movie !!
Stephen Hawking believed AI will threaten the existence of humanity itself, that if allowed to replicate/improve itself, it will eventually take over. (Terminator, anyone?).
Elon Musk called AI a “fundamental existential risk for human civilization”, that it should be regulated now instead of after it becomes a problem.
AI has its uses, but ultimately it’s frightening, IMO. There is an argument to be made, however, that we’ll do ourselves in before AI becomes sentient and all-powerful. Maybe that’s why the real reason there are no people in Spielberg’s movie. I’ve seen it several times. It is fairly schmaltzy but I like it.
And, you’re absolutely right. In 2001, I sat through A.I., left feeling I’d been had, but knowing that the whole thing was a fabulous disaster. Shoulda been shorter, shoulda been more Stanley and less Steven and before I had a blog, I posted a review online. Found that review a few years ago and added it to my current blog in the wake of reviewing Ready Player One.
https://joejots.blog/2018/05/07/spielbergs-a-i/