Talk: Artificial Intelligence

Are the Geth true AI? Legion says they're just software like VIs who interact with each other to form an AI but there's no bluebox involved and they constantly transmit each other through communication lines. 188.27.195.221 23:45, February 11, 2010 (UTC)

Tali talks about this a little in Mass Effect. She says the geth weren't true AI at the start, but her people underestimated the power of the neural network. Matt 2108 23:52, February 11, 2010 (UTC)

Which means that by definition they are closer to highly adaptive rogue VIs than a true AI. Talking with Legion has a great deal more in common with talking to Avina than EDI. Ryvaken 14:25, April 12, 2010 (UTC)

:I would disagree, they are "true" AIs, they just achieve sentience through alternative means. While "traditional" AIs are super-complex single entities achieving sentience by themselves, the Geth achieve the same level of sentience through a collective consciousness. Think of it like neurons in the human brain, alone a neuron is good for nothing, but when combined they create the sentience we enjoy. The Geth operate in a similar manor, only each "neuron" or in the case of the Geth, each "programme" can operate independently to an certain extent. It makes their "thought patterns" or the way they "think" differ drastically from "traditional AIs" but no more different I would imagine as the way a "traditional AI" thinks when compared to an organic mind such as ours. They are just a different type of AI, but no less advanced or sentient. --Looq 20:06, June 21, 2010 (UTC)

The VI versus AI in Mass Effect is also known as weak AI versus strong AI. The VI description pretty much nails weak AI: a computer system of amazing complexity but ultimately nothing more than an interface for a processor, incapable of thought. The strong AI is functionally equivilent to a human mind. Legion and the geth exhibit weak AI traits -- while they are capable of processing new stimuli and able to question philosophical traits in a way that resembles self-awareness, their lack of empathy for others and inability to judge the actions of groups outside their own are classic of a weak AI placed in a position it was never meant to handle. Legion himself, ask it about your N7 armor, it can't answer the question outside electronic parameters. The intuitive leap to nostalgia, hero-worship, or respect is beyond it, even as it mimics that behavior. Talking with Legion, it is impossible to forget that it is a machine. Talk with EDI, Soverign, Harbinger, or that AI on the Presidium, and there could be a life behind their voice. Especially with EDI. In short, the geth are not neurons in a brain. They are switches in a circuit. Ryvaken 12:21, July 17, 2010 (UTC)

I'd have to disagree, at least to an extent. You say that VI vs AI is also known as weak AI vs strong AI, and this is not the case. After all, a weak AI is still an AI, and a VI is not an AI, whether weak or strong. A weak AI is still self aware and capable of independent though, and assuredly is much more than a processor. Basically, what I'm getting at here is that the statements that VI = weak AI, and weak AI = no more than an interface incapable of thought, are simply not true. SpartHawg948 20:53, July 17, 2010 (UTC)

You misunderstand what the terms "weak AI" and "strong AI" mean. They don't mean "dumb true AI" and "smart true AI"; they're actual technical terms in the field of AI. Only strong AI is truly intelligent. A weak AI is one that only poorly mimics intelligence and/or can only behave intelligently in a single specific field. Siri is an example of weak AI, and is very comparable to the VIs we see in Mass Effect. Weak AI is not self-aware in any way. To contrast, Legion and EDI would be examples of strong AI (and also the Council's definition of "AI"). The grandparent poster is absolutely correct that the Mass Effect terms "VI" and "AI" correspond to the real-life terms "weak AI" and "strong AI". --209.222.23.62 02:36, February 26, 2016 (UTC)

Although, I forgot to add: the definition of AI/VI and strong/weak AI has nothing to do with the ability to show emotion, which is why Legion would be classified as a strong AI. He can learn to perform almost any task that an organic sapient being would, given the time. Whether the AI thinks/feels the same way as an organic is completely orthogonal to its intelligence. While some AI researchers see human-like thought as a goal, others point out that this would also mean burdening the AI with all the same flaws as human thinking suffers from (see any list of fallacies). There are pros and cons to human-like thinking, but the fact that a synthetic being thinks differently doesn't instantly mean that it's any less intelligent than we are. --209.222.23.62 02:52, February 26, 2016 (UTC)

Hardware

Do the quality of an AI depends on hardware? Imagine there is a blue box with 1000 processors. the AI is as intelligent as EDI, for example. So if I pick out 950 processors, will the AI be as dump as Barbrady from South Park?

From this article, it looks like the 'blue boxes' are all broadly similar (at least hardware wise). "Without its blue box, an AI is no more than data files. Loading these files into a new blue box will create a new personality, as variations in the quantum hardware and runtime results create unpredictable variations." Besides this, we've never really told that EDI is extra-special for an AI - it's just that she's one of the only AIs we've seen in the game (other than the Reapers and geth, who are both special for obvious reasons). Until we're told what a blue box is exactly, it's hard to speculate about what effect the hardware has on an AI's personality and intelligence. Bronzey 08:28, July 10, 2010 (UTC)

So we don't even know if Blue boxes contain processors? Anyway i mentioned EDI as an "every-day" AI, Just for comparison. May the personality and intelligence of an AI depends on it's blue box. When you copy it to a new, the files will adapt to that Blue box (B.B. in further comments) and that's why they will become new AIs from the data of the older.

Part of Mass Effect's charm is that they don't try to explain technology that doesn't exist yet. But we can make some educated guesses. A quantom computer would, obviously, operate on a quantom level using the forces of the sub-sub atomic, rather than on a molecular level with the forces of electrons as in traditional computing. For our purposes the critical difference is that the quantom level throws out Newtonian physics and creates probabilities, variances, and unpredictable behavior. Thus a "blue box" is by definition unique -- the laws of physics demand that it be all but impossible to precisely duplicate an existing box and literally impossible to measure the duplicate against the original to know if it is in fact a duplication. Because a quantom computer's operation would depend on these quantom variances, no two blue boxes would process the same information the same way. In this manner, the blue box mimics the uniqueness of an organic mind. Ryvaken 12:21, July 17, 2010 (UTC)

Now I undestand that quantom thing.

Eliza

I am more curious about this Eliza AI (or synthetic intelligence if you prefer). It/she is mentioned in the codex entry for Gagarin Station, but it doesn't elaborate on what happened to it/her after it achieved sapience, if the rest of the galaxy knows about Eliza, or even if Eliza is still around D:

We don't know very much about it apart from the Codex entry, and that's about it. Lancer1289 13:56, April 24, 2011 (UTC)

Eva Coré

Video logs on the Cronos station imply she was meant to replace EDI, but I didn't get if she was an AI herself. Was she? Will-O-Wisp

We are currently unsure on this point. Lancer1289 16:43, March 11, 2012 (UTC)