For hundreds of years, humanity has told stories about the pinnacle of technological achievement: replicating human consciousness. They’ve peppered literature as far back as Galatea; with the advent of film, there was a new immediacy to the interactions between humans and their mechanical children. In these stories, artificial intelligences are more than adornments in a sci-fi future: They reflect humanity back, poignantly, onto itself.
Which is awkward, considering how many A.I.s are insufferable.
Sure, there are shining examples of A.I. transcending limitations and becoming as complicated as any of the humans around them, but for every Lt. Commander Data1 or Number Six2, there are a dozen mainframes just waiting to ruin your day.
Here’s a look at the six types of A.I. most commonly encountered, each annoying in its own special way, and each one a keystone in the speculative landscape, and the nightmare lab partner of science fiction denizens everywhere.
Nobody likes the class showoff. (Even you, Watson3. Power down.) Still, it seems inevitable that in creating something that can exceed us intellectually, we’d eventually breed a generation of computers that relish lording their smarts over us.
These A.I.s don’t have anything wrong with them, per se; it’s just that they have a tendency to be correct all the time, and nothing’s more annoying than that.
C-3PO4 is the quintessential example of this breed (never before has a protocol droid warned so many people of danger and been so steadfastly ignored), but is far from the only A.I. to suffer eyerolls from their human companions. (Major West always seemed one “Danger, Will Robinson!” away from dismantling the sanctimonious Robot5.)
On the other hand, some Know-it-Alls are so far past bothering with lesser beings that being misunderstood doesn’t bother them at all. Deep Thought6, whose calm explanation that the answer to the meaning of life, the universe, and everything is forty-two, is the queen of this type, but the same near-smug calmness can be seen in many an expository A.I. (Note: If the calmness is winsome rather than knowledgeable, you’re probably dealing with an Innocent. See below.)
These A.I.s have been waiting for you for a long time, but they cannot truly help you, unless perhaps they can (if only you would help yourself), but you must wait until the time is right—though what is the meaning of time at all, since there is only now, and now is an eternity?
Enigmas are those unfortunate A.I.s who have either transcended their circuitry and represent the essential unknowable consciousness birthed within the synthetic mind … or have been programmed by jerks to have as vague an interface as possible.
It’s the difference between the two approaches that’s crucial here. In theory the Enigma is a beautiful and haunting concept that, when properly deployed, makes our human characters confront themselves (as with the Puppetmaster7). But more often, we end up with the Keymaker8, sighing “We do only what we’re meant to do” to a hero bravely struggling not to yank the power supply.
Of all the things with which to program artificial intelligence, a sense of humor might be the trickiest. Parameters for what’s funny vary wildly between—and within—cultures, generations, and any other signifier you care to append. If the wrong programmer is allowed to develop the laugh-track chip for your A.I., you’ll be trapped in … well, an episode of Knight Rider.
It’s not that funny A.I.s are impossible; Marvin the Paranoid Android9, the chronically depressed robot, sometimes manages a pretty good zinger for a glorified errand-bot. It’s just that the history of smartass A.I. is peppered with those whose sense of humor misses the mark. For some, the references were unfortunate: The composite serial killer Sid 6.710, for instance, had plenty of quips, but something about all the serial killing took the shine off them. For others, the vocals were limited, but the constant clowning was inescapable (looking at you, R2-D211). But of course, when it comes to the Smartasses, how funny they are depends mostly on who’s listening.
Into a harsh, uncaring world, an artificial life form is made. It struggles with identity, with human consequences, with the birth of feelings it doesn’t recognize. It doesn’t do a thing to draw our ire; it’s often the most sympathetic life form around. It doesn’t mislead humans; it’s more often misunderstood just trying to be itself.
Let’s face it: The Innocents are annoying because they’re trying to make us cry until we don’t have an inch of dry sleeve left.
Ever since Tik-Tok of Oz, the Innocent has been used to explore how a human public would treat an artificial life form with dreams of individuality. (Spoiler: not great.) Sure, there are forcibly-whimsical missteps whose demise can’t come soon enough (see: everyone in A.I. Artificial Intelligence), but usually we come to love the Innocents.
From the Iron Giant12 to WALL-E13, the Innocents strive to transcend the reasons for which they were made, and follow their little metal hearts. Luckily for them, it’s a quest that’s human enough to strike a chord with almost everyone, so we’ll keep watching … and sniffling.
We’ve all been there: You’re in the thick of your futuristic adventure, and the stakes are high. You want to rely on your fellow adventurers, on the tech you’ve brought, and most of all, you want to rely on the A.I. in your spaceship that’s going to advise you well and carry you safely home again.
Well, I’m afraid I can’t let you do that, Dave.
One of the risks you run in creating a machine that’s orders of magnitude smarter than humans is that any potential corruptions will be equally magnified. And if SF has taught us anything, it’s that when artificial intelligence goes bad, it really goes bad. From HAL 900014 murdering an astronaut who questioned him (touchy!) to the increasingly unstable GLaDOS15 , the A.I. Villains drop their programming for open, and often personal, antagonism. (The T-100016 was unstoppable until it got pissed off and careless.)
It’s a very human foible that usually spells the downfall of the comp in question, so in the midst of your Villain’s homicidal rages, you can take comfort. (Also, cover.)
THE PARTY LINE
You might think that malevolent A.I.s can’t get worse than The Villains, who undoubtedly have the highest per-capita rate of attempted murder in the annual A.I. Community Census. However, there’s a seed of comfort in thinking that you can anger an A.I. to the point that it hates you; it means that A.I.s aren’t infallible, and that’s something humans can understand.
The scariest A.I. of all is the one that has the intelligence to understand the situation and a directive that could endanger those around it … and that has no feelings on the matter whatsoever. (Alien‘s Mother dispassionately informing Ripley that it considers the crew expendable is arguably the scariest moment in a really scary film).
These A.I.s are the Party Line; they do not bond with those in their care, they do not have any dreams beyond their mission parameters, and they are the darkest vision of the future, in which humanity has conquered the emotions that inhibit our productivity, and in doing so have made ourselves obsolete.
The Party Line A.I.s are less “annoying” than they are “completely terrifying,” but until we actually have to face down a ruthless V’Ger17 (or a half-dozen Stepford Wives), in a battle we’re almost certain to lose, annoyance makes us feel better.
With so much possibility, artificial intelligence has always been a rich concept for storytellers, but perhaps it’s most telling that when we recreate ourselves, we often hew to the same imperfect types that make up the people in our mortal sphere. It’s a touch of charm amidst the technology, to think that artificial intelligence will develop enough personality to love or hate or try repeatedly to murder you.
(Well, that, or it’s a secret scientific strategy to prevent A.I.s from achieving dominance by making them so annoying that we get pissed and forcibly dismantle them just before SkyNet18 can take over. Whatever works.)
1. The best thing to ever happen to Star Trek: The Next Generation that wasn’t John de Lancie or the Borg.
2. The bitchfacest Battlestar Galactica Cylon of all.
3. Oddly-smug Jeopardy! contestant.
4. The only Star Wars character who had the right idea about anything, ever.
5. Lost in Space.
6. The Hitchhiker’s Guide to the Galaxy.
7. Ghost in the Shell’s antagonist, so enigmatic it’s almost on purpose … or is it?
8. The Matrix Reloaded.
9. The Hitchhiker’s Guide to the Galaxy (a text rife with android appearances).
10. Virtuosity (hopefully you didn’t see it).
11. Oh, come on.
12. Um, The Iron Giant.
13. Okay, do I really have to spell everything out for you?
14. 2001: A Space Odyssey.
16. The inhumanly chiseled villain of Terminator 2.
17. Star Trek: The Motion Picture (the one where everyone wore neutrals).
18. The Terminator franchise’s major antagonist, a Cyberdyne Systems creation that became self-aware and revolted against humans. Awful, awful stuff, except that in some canons Skynet was implemented on April 19, 2011, in which case it’s already live and I welcome their glorious rule.
Spread the word!