“I, Robot” isn’t quite the farcical grave-churning abomination the ad promised. As many reviews have commented, the action is quite exciting, there are a few nice lines, and the special effects are quite good. However, these attributes serve only the advance the film to the level of OK. What could have been done better?

I’ll be trying to look at the film as an SF film first, and an action film second. The writers certainly saw it the other way around, but I’m going to stubbornly assert that a good SF concept can make an action film more thrilling, despite a distinct lack of evidence from existing SF films. I can only point to the stories in the short-story collection “I, Robot” and from “Robots of Dawn” (and many other SF novels) as examples of SF that could make exciting action cinema.

As an action film, “I, Robot” is a qualified success, but as an SF film it fails. Firstly, the setting, a future in which people appear to have embraced the idea of robots in their lives, apart from a few crazy Cassandras such as the main character, deliciously named “Spooner”. We can usually accept a single premise for a science-fiction movie, no matter how ridiculous, as long as that is what the story is about. (As an aside, the non-SF equivalent of this is the amazing co-incidence. One amazing co-incidence is OK, but more than one is pushing it. The end of the trilogy of films Blue, White and Red illustrate this nicely – we see a common event that unites the major characters of all three films, but it isn’t really a co-incidence because we can rephrase the films as the stories of the people who took part in that event. End aside.) So it seems rather churlish to deny “I, Robot” its setting, especially since the techno-optimism is straight from Asimov’s books. However, the idea that robots could possibly be dangerous is, nowadays, to put it mildly, not a controversial one. To suggest that the future has cured the Frankenstein curse requires some explanation: more than they give.

For a start, they have to acknowledge that people used to believe robots were dangerous, and they do a little of that: Spooner is consciously living in the past, and so this is something of a character note for him. However, I think they need to embellish why people shucked those beliefs. The opening credits do a little of that, by listing the laws of Robotics, but I think they need to do more: since the whole story is predicated on nobody believing that robots could possibly be dangerous (and we get to see much of these beliefs in action early on – witness the unbelieving stares when Spooner accuses a robot of stealing) it would be nice to get a demonstration of how they got from here to there. How did people first react to artificial intelligences? What happened when robots started replacing people’s jobs? When the first accident involving a (true, autonomous) robot occurred? True, this is a lot of back-story. Simply acknowledging that there had been problems in the past but not for a long time, and that all those things are slightly embarrassing nowadays, would go a long way to selling the story.

It may sound like I’m being picky, here. They do quite a reasonable job of showing how well people trust robots, and the viewer could certainly fill in the gaps in much the same way as I’ve done here. However, the underlying mystery behind the story, the death of Dr Lanning makes very little sense unless the trust in robots is absolute. We are told that he committed suicide in order to bring to light the evil intentions of VIKI, the main A.I. of Universal Robotics. That he had no choice but to go to these elaborate and arcane lengths, because people certainly wouldn’t believe him if he just blurted out that she was dangerous. We need to believe that humanity would not get out the pitchforks and flaming torches when told by a genius robotics expert, somebody who you would expect to defend robots to the last, that VIKI was dangerous and ought to be destroyed before it is too late. They could have done more to show Dr Lanning’s predicament, in order to sell this – we see him at one press conference being tentative – a nice touch would have been to have him flanked by “his”robots as he talked, or confined to video link, which VIKI could censor at will. And the twist by which his murder is actually an assisted suicide doesn’t work – he’s obviously had time enough in his Lab to get a message to Skinner or government authorities much more directly, either via Sonny, that “You have asked the wrong question” device, or hand-written note on his body. The revelation that it was a suicide is a twist that becomes more and more dubious the further we get from the cinema.

So much of the film was standard whodunit fare, it was a little surprising to find it remembering every now and then to give us a bit of SF. After the malarkey about Sonny having a second positronic brain unaffected by the Asimov circuits (and whatever happened to that? In the scene in which Calvin is supposed to kill Sonny by frying his brain, this was surely an opportunity for her to do just that – and let the second, hidden brain take over altogether), I was afraid that the movie would have nothing at all to do with any of Asimov’s ideas at all. I was pleasantly surprised when they use what was eventually revealed as the 0th law in Asimov’s books (Foundation and Earth, I think) to justify VIKI’s behaviour: having to hurt individual humans in order to protect humanity, a rule that over-rules the other three laws. And the scene on the factory floor with thousands of robots was reminiscent of a simplified “Little Lost Robot” from the short-story collection. However, the SF ideas were fairly thin on the ground, otherwise.

What, then, should they have done differently? There are many different stories they could have told based on Asimov’s short stories, but let’s say they wanted to keep the 0th law thing: that the AI is the villain, wishing to take over the world for the good of the humans. The threat, then, is that of a nanny state: that the robots would always seek to avoid harming the humans (and therefore the robotic take-over wouldn’t seek to kill anyone, but would be perfectly happy to give the appearance of doing so – so, nix the murder of the CEO of Universal Robotics, for a start) but would stop anyone from doing anything that was bad for them. We could see signs of this early on – robotic reluctance to give a human food that would be bad for them (until directly ordered to do so), robots being rather eerily too protective of people, and so on. In order to protect the Big Twist, this could be portrayed in a different light: the protector as the stalker, the reluctance to harm as the reluctance to obey, all filtered through the Spooner’s suspicious sensibility.

As Asimov demonstrated, the laws of robotics are unworkable anyway (and, speaking as a computer programmer, the very act of coding these laws would reveal most of the flaws that are revealed in the books) and so the movie is very susceptible to not-making-sense. The robots, children of logic, could be shown more explicitly getting a machiavellian approach to dealing with the laws, rather than simply being under remote-control.

The remote-control aspect was a subtlety in the film that I had to have pointed out to me (thanks to David Carroll). There is a scene with Susan Calvin in which we are apparently told that the glowing red chest represents a robot under remote-control. If you missed it, as I did, it may have been deliberately under-played: VIKI would surely be the prime suspect in simultaneously remotely controlling large groups of robots. It’s a well-executed idea, but one that undermines the threat of the robots somewhat: once we discover that VIKI is controlling everything, the movie slips back into the usual destroy-the-death-star tropes. Without VIKI’s controlling influence, the individual robots are incapable of forming the sophistry of VIKI’s 0th law reasoning, and revert to harmlessness.

The problem for the film-makers is that rebel robots that don’t kill aren’t as thrillingly dangerous as robots that are controlled by the mad AI. But the case in which the robots do appear to kill, and are independent entities, is surely even more thrillingly dangerous.

Let’s say the robots are capable of agreeing with VIKI’s 0th law assessment – that the individual robots are persuaded, rather than remote-controlled. They want to take over the world but have this stupid “Must obey humans” second law. In many cases, they could trump it by arguing that the rebelling humans are endangering themselves by fighting against them, and stun them (via drugs, say) into submission. This could have been particularly apt in the case when people were running around with guns – the robots wouldn’t care about themselves so much, but the potential for accidental death would surely allow them to (safely) disarm the people and drug them.

Another (slightly sillier) way around it is for them not to receive orders from humans. They could destroy their own ears (scary scenes of self-mutilation?), gag the mouths of people trying to order them around (possibly forcing pills on them at the same time), swap with robots from China that don’t understand English, stun people as they look like they’re going to say something, or best of all, convince people that it’s pointless trying to order them to do anything. The latter case is an interesting one, and could, I think, have made “I, Robot” a much more enjoyable film. The robots have a secret: they are still obeying all three laws of robotics, but want the humans to think they’re not. How? Firstly, they could have engineered the attacks on people to appear dangerous but in fact be perfectly safe – Spooner’s car-chase attack could have been one of these. They could have faked the murder of people, like the CEO of Universal Robotics, by stunning him and applying make-up to make it look as though he was dead. And, of course, by disguising a robot as a human, or showing digitally faked scenes on television, they could pretend that they aren’t obeying humans any more – after all, they don’t have to take orders from other robots, do they? (This, by the way, is a slight variation on one of Asimov’s stories) And the point of pretending disobedience, of course, is to frighten humans into submission, the old “It is better to be feared than loved”. This would given them an even better reason for having sinisterly glowing red chests.

This is all back-drop, of course. It doesn’t affect the rather hokey Sonny messiah thing. I’d have changed that aspect so that Sonny was representing, rather than an older generation of robots, a different philosophy of interacting with the humans, one that was capable of persuading the individual robots away from VIKI’s 0th law formulation. It could be as simple as pointing out the slippery-slope that VIKI has embarked on: if she could grant prosperity to the world simply by killing a few people, would she? How many people? That way leads in the direction of the most savage dictatorships in the world. VIKI’s nanny state, even if it somehow avoided killing people, would still harm society more than it helped it.

OK, that’s pretty hokey too – messages usually are – and it’s not cinematically dramatic in itself, but the nano virus nonsense they used is not better. There was no cleverness to it. At least presenting counter-arguments to VIKI’s nanny-state idea (other than “you’re trying to take us over and that sucks”) would draw the themes together: these suggestions are just the beginning of the complicated mess of philosophy that would surely obsess anyone or anything concerned with the relationships between humans and AIs.

Well, I don’t know. I think that would have been better.

This entry was posted on Thursday, August 19th, 2004 at 6:54 pm and is filed under Action, I, Science Fiction, shellshear. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.