Friday, June 22, 2018

A Robot Learns to Love Itself: Reflecting on the Murderbot Diaries by Martha Wells

The Murderbot Diaries is an AI self-actualisation story which takes us far beyond the basic "can a robot feel?" question that is still the standard starting point for these kinds of tales.


There's a moment near the start of Rogue Protocol, the third in Martha Wells' Murderbot Diaries series (forthcoming August 7, 2018 from Tor.com Publishing), that quietly broke my heart. The self-proclaimed Murderbot, a rogue SecUnit (a human-robot hybrid "construct") which hacked its own governor module after an unfortunate murder-based incident that was subsequently wiped from its memory, is trying to distract itself from the endless, stupid problems of humans by watching a new show. Unfortunately, the plot isn't working out, and Murderbot is eager to get within range of a station so it can download something different. If only, it tells us, this terraforming horror series had a rogue SecUnit character who could stop the squishy humans from all getting horribly killed...

On the surface, this doesn't seem like a big deal. Murderbot watches rather a lot of shows  indeed, extensive media consumption is its most prominent character quirk  and it also does a lot of complaining, so the combination of the two is not exactly unusual. However, this is the first time it has articulated a desire to see itself represented positively in media. In the previous book, Artificial Condition, Murderbot had explained to its new "friend" ART the Asshole Research Transport (long story) why their favourite TV show is Sanctuary Moon, a show in a setting with no SecUnits or security issues at all. ART's favourite shows, in contrast, all tend to involve spaceships protecting their humans. Rogue SecUnits in media are all portrayed terrible monsters, because, it thinks, that's a rational way of looking at rogue SecUnits in general. To even fantasise about the existence of a heroic rogue SecUnit one book later is a serious step forward for Murderbot, even if it doesn't acknowledge the change of heart itself.

It's this constant grappling with the character's identity and self-worth that really lifts the Murderbot Diaries (a series which began with last year's Nebula winning, Hugo nominated All Systems Red, and is due to wrap up later this year with Exit Strategy) from being a merely hilarious story about a cynical construct to being something rather special. Like Moon, the central character in The Books of the Raksura (Martha Wells' other Hugo finalist this year), Murderbot is a convincingly non-human person who blends recognisable emotional responses with occasionally very alien reactions and behaviours; both are outsiders who find themselves offered friendship and community but have to learn how to accept it. In telling story it does, the Murderbot Diaries also turns the traditional robot narrative on its head: Murderbot isn't a robot learning to feel, it's a robot who is already all but overwhelmed by its emotions and has to learn how to manage and express them in a galaxy where many people still treat it as an unthinking tool.

And while Murderbot has it worse than most, it's apparent that a lot of people in this universe  be they humans, bots or something in between  are similarly struggling to establish their right to live and flourish beyond their usefulness to all-powerful corporations, who are not above mass murder to get what they want. When we first meet Murderbot in All Systems Red, it's been hired out to an uninhabited planet with a group of naive but (it grudgingly admits) likeable humans who are conducting a survey, when they find themselves in the middle of a highly subtle corporate assassination attempt. Murderbot, who has already gone rogue by this point but is pretending to be compliant, ends up accidentally "outing" itself as a fully realised sentient when it has to evacuate an injured party member, and spends the rest of the mission attempting to rebuff attempts  particularly from the mission leader, Dr. Mensah  to talk to it about its feelings and treat it like a person.

Murderbot is quick to tell us that this is because it doesn't want to make the humans uncomfortable, and the reader just as quickly realises that this is a planet-sized act of projection on our hero's part. Faced with a group of people with no preconceived notions of what a SecUnit should be, who discover that it's not a heavily armoured machine but a being with a human face, the ability to conduct caring small talk, and a massive addiction to trashy soap operas, means Murderbot suddenly has to cope with being treated like a person, forced to earn trust and friendship from its coworkers and treated accordingly when it does. To someone who has thus far dealt with being emotionally sensitive by insisting that nobody cares and hiding itself behind an opaque visor, this change is nearly impossible to process.

Because behind the sarcastic asides and wry commentary, Murderbot's narration is a veritable bingo sheet of unhelpful thinking styles; its propensity to internal self-sabotage is both relateable and excruciating to watch. Everything Murderbot does right is disqualified from positive consideration because it's just what SecUnits do, while everything that goes wrong is a total disaster that's all its fault. All of Murderbot's strengths are flukes or basic programming, while its weaknesses are all-consuming. Because Murderbot is very anxious around people, people must be objectively difficult things (except in media, where they are fascinating and enjoyable). All Rogue SecUnits are terrifying, terrible individuals who are very rightly the villains whenever they appear in media, and would be awful to meet in real life. Oh, and of course it's labelled itself Murderbot (and the first bot who sits down to watch TV with it "Asshole Research Transport"). While our hero does indeed recognise and label its own thinking as "anxiety", and can demonstrably think things through or talk itself down when required, the narration doesn't give us much second-order thought or self-reflection, leaving the mechanics of growth behind the scenes and leaving us with only subtle signs of growth behind Murderbot's aggressively curated self-image.

Later instalments have Murderbot truly going rogue and, in the process, straying even further outside of its comfort zone, passing as human while it pieces together evidence against the company which attacked its humans and discovers more about its own past (including the event which led it to call itself "Murderbot" in the first place). Intentionally or otherwise, it finds itself spending more time with humans similar to Dr Mensah's group: people it ostensibly finds insufferable, naive and incapable of staying out of danger but who treat it like a person, even when the "augmented human" identity slips.

We also get interactions between Murderbot and other bots and constructs, most notably ART in Artificial Condition, and the irrepressibly friendly (and, apparently, extremely annoying) Miki the helper bot in Rogue Protocol. Murderbot is rather rude about both of these characters, especially Miki, who it dismisses as a "human's pet": a dismissal which likely reflects its feelings about being offered a similar choice earlier in the series, rather than being directly Miki's fault. However, even while it's calling its fellow bots assholes and pets, Murderbot is also completely willing to accept them as people and in many ways treats them the same as humans: trustworthy in some ways but likely to betray you when their "programming" requires it. Even bots with demonstrably low capability get treated with respect by Murderbot, although it always puts its own self-preservation first. We are led to suspect the only thing that isn't a person to Murderbot is Murderbot itself: an ironic conclusion for the character to arrive at, given its narration leaves the reader in no doubt that Murderbot is quite definitely "one of us".

The Murderbot Diaries is an AI self-actualisation story which takes us far beyond the basic "can a robot feel?" question that is still the standard starting point for these kinds of tales in SFF. The series presents us with a robot character who we immediately accept as a funny, cynical, highly competent and resourceful person, and who I suspect many of us would love to hang out with, even knowing it would probably complain internally and make up rude nicknames for us if we did. In doing so, The Murderbot Diaries gives itself room to ask more complex questions about the relationship between how we see our own personhood and self-worth compared to how others see us; and how to find healing, growth and self-expression even when all one wants to do is self-isolate. For Murderbot, it's a slow, frustrating journey, and one which is largely obscured by bluster and sarcasm. But when the moments of growth shine through  when the Murderbot accepts that it might just be hero material  it's are all the more poignant and exciting for being so hard-won.

POSTED BY: Adri is a semi-aquatic migratory mammal most often found in the UK. She has many opinions about SFF books, and is also partial to gaming, baking, interacting with dogs, and Asian-style karaoke.


References: Wells, Martha. All Systems Red [Tor.com Publishing, 2017].
                    Wells, Martha. Artificial Condition [Tor.com Publishing 2018].
                    Wells, Martha. Rogue Protocol [Tor.com Publishing 2018]