The convoluted logic of the Mass Effect trilogy's controversial ending hinges on the idea that sufficiently advanced species will inevitably create artificially intelligent life that will rebel and, if left unchecked, exterminate all organic life in the galaxy. To combat this threat, the Reapers harvest advanced civilizations, giving primitive ones the chance to flourish without being snuffed out in their infancy. This account of the Reapers' solution blindsided many players because it placed one of Mass Effect's weakest themes at the core of its most important conflict. The Mass Effect games never coherently convey the impression that synthetic intelligences pose a unique threat to all life.
Although killer robots have been a staple of science fiction for some time, this is actually a hard case to make, especially in a fictional universe with superluminal travel. Synthetic intelligences do not require a comfortable atmosphere or gravity, and can function at a wider range of temperatures and radiation levels than can humans. As a result, robots that reach consciousness have no particular need of the star systems organic lifeforms inhabit. They can happily occupy any bright (for energy) star with some metal-rich terrestrials and asteroids (for resources). Unlike organic life, they will have no intrinsic imperative to reproduce, limiting their need for expansion. Even if they do grow, the number of star systems that can support synthetic life is likely to be so vastly greater than the number that can support organics that resource exhaustion and subsequent conflict is unlikely to occur for millennia.
Considered seriously, artificial intelligences pose little threat to organic life, significantly less than interspecies conflict (i.e. the Rachni) or unintelligent tech-life such as grey goo. The inherent illogic of this theme means that it must be sustained by knee-jerk fear of the Other, and by direct demonstrations of the threat of synthetic life in the game world. The grand narrative of the Mass Effect trilogy involves so many alliances with alien races that the first factor cannot seriously contribute. Even the games' characters seem ambivalent on the otherness of synthetic life, as this conversation from Mass Effect 3 (recorded by Krystian Majewski) attests:
This means that the universe must lean on its existing artificial intelligences to convey the threat. Commander Shepard encounters three groups of synthetic intelligences: the Geth, rogue programs, and the Reapers themselves. Of these, the Geth receive the most comprehensive exposure. From the game's first moment, these networked intelligences are presented as enemies, slaughtering much of a human colony and serving as the principal enemy force throughout the first game.
Putting the Geth on the business end of the protagonist's gun adequately serves the theme, but the Geth never manage to make a case for themselves as a catastrophic threat. In part this is because their story cannot be separated from that of their creators, the Quarians, who have been forced to flee their home systems and now live in a fleet of starships. Perhaps this would not sound so familiar were the games not contemporaneous with the astonishing re-imagining of Battlestar Galactica. In its presence, however, the Geth and the Quarians came across as a cliche, something not to be taken seriously.
Mass Effect only separated itself from its obvious inspiration in that the Quarians ultimately turned out to be the villains. From the very first conversation with Tali aboard the Normandy, Shepard can point out that the Quarians tried to kill the Geth first. In the final game, data the player can find on the Quarian homeworld shows that the Geth only ever fought their creators in self-defense. They never rebelled; they were attacked.
Even the hostility of the Geth from the first game gets walked back. Mass Effect 2 introduces a sympathetic Geth teammate named Legion, who explains that the inimical Geth from the first game served the Reapers in hopes of learning how to ascend to a higher level of intelligence. The synthetics who fought Shepard throughout the first game did so not out of any intrinsic desire to exterminate organic life, but rather as mercenaries. This puts them, at best, on a level with the Vorcha or Krogan in terms of their danger. In the final game of the trilogy, the Geth ally with the Reapers again, but even this isn't because they want to attack organics. Rather, they turn to the Reapers as a matter of self-preservation after a successful attack by the Quarian fleet.
Shepard comes into conflict with the Geth throughout the Mass Effect series, yet these encounters never make a coherent argument for an inevitable extermination of organic life by synthetic. Each fight with the Geth arises because of an attack on them or the external stimulus of the Reapers. When the Geth-Quarian conflict comes to a head in Mass Effect 3, the player can choose to broker peace between the warring parties, contradicting the supposed theme of conflict completely. Nothing in the Geth storyline coherently communicates the idea that synthetic intelligences are inherently dangerous to organic life.
A somewhat less sympathetic foe comes in the forms of rogue programs. Illegally-constructed AIs or rebellious virtual intelligences (VIs) threaten Shepard's safety fairly regularly, especially in the first game. Yet, with few exceptions, these actions are defensive. The rogue VI on Luna and the illegal AI on the Citadel are just trying to stay alive. The human core of the "rogue VI" in the "Overlord" add-on for Mass Effect 2 has been tortured to the point of insanity. Very few of these rogue programs are acting out of considered aggression.
The series undercuts the threat of rogue programs more spectacularly through the character EDI, an AI that operates many of the systems in the Normandy. Not only does EDI prove extremely helpful throughout the two later games, she can even form a romantic relationship with the ship's pilot, Joker. Late in Mass Effect 3 it is revealed that she was reconstructed from the rogue programs on Luna and the Citadel, completely transforming them from foe to ally.
While the rhetoric of gameplay, especially in the first game, positions these synthetic intelligences as enemies, the narrative components of the games argue that they are innocent, or even helpful. In Mass Effect 3, even the gameplay angle falters, as EDI can enter the field as the player's ally in combat. Again, the game's message is mixed, and fails to effectively argue that synthetic intelligences are a unique danger.
The only synthetic foes that seem to present an unalloyed threat in the Mass Effect series are the Reapers themselves. Even that gets moderated in the finale, when the Catalyst reveals that their purpose, however grimly executed, is to preserve the possibility of organic life. The rationale for this campaign to extinguish advanced civilizations, though, requires that AIs other than the Reapers themselves pose a danger. Otherwise, the game legitimately opens itself up to the charge that the Reapers kill people to keep them from getting killed by Reapers.
In this respect, the Mass Effect series fails. Synthetic intelligences clearly pose a danger, but they are an ordinary hazard, a race like any other, that can be defeated or even made into an ally, or a lover. The player reaches the endgame without any sense that synthetics other than the Reapers themselves pose an insuperable threat, and so the explanation given for the Reapers' behavior comes as an inexplicable and deflating surprise. Having drawn the idea that AI poses a threat to organic life from more compelling science fiction universes, Mass Effect undercuts that conceit by adopting an outlook that, if not exactly Asimovian in its optimism, supposes that AI and humans can at least coexist in peace and fruitful collaboration. The B-movie concept of killer robots can't survive nuanced or mature examination, and its collapse makes this key theme one of Mass Effect's weakest.