So I just started re-reading The Fall of Reach for like the 8th time and I noticed that Déjà was being used by Halsey when the Spartans were first conscripted in 2517, and still in use 8 years later in 2525 when they received their augmentations. How did she get past the 7 year lifespan of an AI? Sorry if this has been brought up before.
I believe Deja was a “dumb” AI which allowed it to live longer.
Riiight. Duh, I shoulda known that. Forgot they had a longer lifespan. Thanks!
The 7yr rampancy marker isn’t constrained to exact numerical output. It’s at around the 7yr mark that an AI begins to run out of room to process new information, which then leads to the first or sometimes second stage of rampancy. Depending on how much information the AI is exposed to or willingly seeks out, the actual “end date” may vary slightly. But it’s safe to say that after 7yrs, there’ll be signs of some kind.
Gotcha. I knew the 7 year mark isn’t exact, but I just thought I’d ask. I’m trying to start being a little more active in the forums. I’ve been lurking here for years now so I figured I might as well start.
Just to add, in The Cole Protocol, the AI in use by the humans (colonists?) was definitely over the 7 year mark. In fact it commented that it was probably going rampant. However it was still largely holding all of its marbles.
So the 7 year mark is a guideline, rather than a rule. And how quickly the AI degenerates will be a function of the amount of knowledge it has been exposed to. Given Cortana was hyper exposed (between the fragment on Reach, the info from Halo CE, the data from raiding the various covenant ships . . .), I’m surprised she held on as long as she did.
She wasn’t a Smart AI.
> So the 7 year mark is a guideline, rather than a rule. And how quickly the AI degenerates will be a function of the amount of knowledge it has been exposed to. Given Cortana was hyper exposed (between the fragment on Reach, the info from Halo CE, the data from raiding the various covenant ships . . .), I’m surprised she held on as long as she did.
I believe there was a lot more to Cortana than we’ve been allowed to know for the moment. Her time with the Gravemind and all of her exposure to Forerunner tech- perhaps she had been able to extend herself to some degree, beyond her original programming. She was, however finite, a very advanced AI.
Dumb AI’s likely last longer because they don’t make as many connections between information they have. They’re more like Siri with a personality.
I believe the ONI AI in FoR also lasted a considerable number of years.
Déjà was still active up to the fall of Reach in 2552. It seems “dumb” AIs have a very long life span. Could the Assembly be a collection of dumb AIs? That could explain how they have been around so long. I don’t know of anything they have done that shows if they are “smart.”
> Déjà was still active up to the fall of Reach in 2552. It seems “dumb” AIs have a very long life span. Could the Assembly be a collection of dumb AIs? That could explain how they have been around so long. <mark>I don’t know of anything they have done that shows if they are “smart.”</mark>
Manipulate Humanity for hundreds of years, causing the Spartan project, controlling certain parts of the war with the Covenant, ensuring the survival of their creators.
I don’t think a Dumb AI has ever been seen to have the ability to have such an influence. Their discussions with each other are another smart AI characteristic, they are making decisions entirely on their own, human elements, not bound to a specific program or task as Dumb AIs are used for.
Couldn’t that be their specific task? I know it’s a huge one, but that could be what their creators, whoever they are, programmed them to do. I could be wrong, but didn’t one of the books state that dumb AIs could work together and increase their abilities?
Also, I didn’t think interacting with other AIs was a characteristic marking smart AI. Or do you just mean the breadth of their conversations?