No one is surprised when the advanced AI turns on its creators. At this point in the history of speculative fiction it is almost a given – so to make the villainous reveal satisfying the author needs to have constructed a believable character. Some authors pull this off well, not that their reveals are necessarily surprising, more so that they are satisfying for a reader. On the other hand, there are creators who very much lean into the trope and use it as a shorthand to take away from genuine character building. This shortcut taking is obvious in films like Avengers: Age of Ultron while books like Chuck Wendig’s Wanders make some attempt to mask their AIs in mystery. Convincing stories about rogue technology require character building and an awareness of audience in order to create an immersive believable narrative.
One of the more frustrating things about the way that Ultron’s character is created is the use of JARVIS as a base but not the defining feature of the character. If JARVIS had gone rogue on his own it would be fascinating, the emotional connection to Tony Stark and established personality could present compelling story of an AI who’s twisted sense of moral duty puts it into conflict with its emotional ties to its creator. Instead Marvel sidestepped the emotional weight of JARVIS willingly turning on Tony and had Ultron be a separate, new entity created by Tony that corrupts JARVIS despite the original’s attempts to resist. Age of Ultron is a typical tale of scientific hubris by a charismatic main character which lacks emotional depth that could push the story past a replaying of tropes.
The AI in Chuck Wendig’s 2019 book Wanderers lacks the kind of personality that is present in both the AIs in Age of Ultron. For much of the 800 page book Black Swan doesn’t even speak, communicating through much more traditionally mechanical means while also creating a feeling of near omniscience. Black Swan is very clearly not to be trusted, the reader knows this and so does the protagonist of the novel. Wendig confronts the trope of the rogue AI directly and instead builds up tension as the reader and protagonist jointly wait for inevitable betrayal. It is not necessarily that the book is unconcerned with the reader’s lack of surprise as it is a matter of the author playing on those expectations for maximal impact. The trope is further complicated by the fact that Black Swan never truly turns on Dr. Benjamin Ray, its wielder if not necessarily its creator. Instead the culminating moment is Black Swan revealing its trump cards to prove that it is in control of Dr. Ray, not the other way around. This AI does not want to destroy its wielder and has no interest in using physical force to achieve compliance, rather it has manufactured a situation where its control is absolute.
Having revealed a villainous AI there is then the question of resolution, what are the characters of the story going to do in the face of their creation turned against them? The answer in Age of Ultron is that they are going to beat each other into a bloody pulp. For the Age of Ultron movie this is even a clear-cut heroes vs villains battle against evil robots. There is fall out from Tony’s creation but the majority of it is stuffed into a sequel full of human conflict and little interest in problems surrounding artificial intelligence. In contrast, Wendig continues to build up tensions between Black Swan and its wielders as Dr. Ray has to decide whether to be complicit in Black Swan’s plans which are already too advanced to be stopped. Where Age of Ultron follows in the footsteps of stories like 2001: A Space Odyssey which sets up a humans OR technology dichotomy; Wanderers is a book filled with choices that deliberately undermine that dichotomy.
None of this is to say that Age of Ultron isn’t a ‘good movie’ because many people liked it for a wide variety of reason. Nor is Wanderers some shining paragon of the perfect novel. However, the way in which these two stories handle a narrative surrounding a rogue artificial intelligence demonstrates the challenges of innovation on an established speculative fiction trope. The reader or viewer is probably never going to be truly surprised when the writer reveals that some advanced technology has turned on humanity, so there needs to be other sources of tension. Age of Ultron had an opportunity to build emotional tension around the AI but chose instead to use human conflict (likely because that is what fit better into the already planned cinematic universe). In contrast, Chuck Wendig has given his protagonist the same knowledge of the rogue AI trope that the reader has and then uses that knowledge to force the characters to confront their own choices in the face of a masterfully manipulative new being. As the real world of technological advancement brings AI and algorithms into broader areas of people’s everyday lives I think these kinds of speculative fiction stories will continue to be popular. Speculative fiction often reflects the things that people worry about and explore worst case scenarios. Creating interesting stories about rogue AI requires work to move beyond the replaying of tropes but also offers the opportunity to suggest real questions that are of interest to contemporary readers.