Return to Unfiction unforum
 a.r.g.b.b 
FAQ FAQ   Search Search 
 
Welcome!
New users, PLEASE read these forum guidelines. New posters, SEARCH before posting and read these rules before posting your killer new campaign. New players may also wish to peruse the ARG Player Tutorial.

All users must abide by the Terms of Service.
Website Restoration Project
This archiving project is a collaboration between Unfiction and Sean Stacey (SpaceBass), Brian Enigma (BrianEnigma), and Laura E. Hall (lehall) with
the Center for Immersive Arts.
Announcements
This is a static snapshot of the
Unfiction forums, as of
July 23, 2017.
This site is intended as an archive to chronicle the history of Alternate Reality Games.
  

The time now is Sat Nov 23, 2024 5:20 am
All times are UTC - 4 (DST in action)		 View posts in this forum since last visit
View unanswered posts in this forum
Calendar
 Forum index » Archive » Archive: The Haunted Apiary (Let Op!) » The Haunted Apiary (Let Op!): General/Updates
[SPEC, LONG] AIs, Reflection, and Cupid's Knife
View previous topicView next topic
Page 1 of 2 [25 Posts]   Goto page: 1, 2 Next
Author Message
BeeNetter
Boot


Joined: 13 Oct 2004
Posts: 49

 		[SPEC, LONG] AIs, Reflection, and Cupid's Knife
I've been lurking for a long time (thus my avatar) and this is my first real post. Warning: It's a LONG one.

AIs in the Halo universe are extremely powerful. When dealing with Shibura, Durga implies that she could "cook him" or "send him to jail". The Operator's Monologue confirms this - Melissa's searches for "central thermo controls", "slaved AI", "vehicle controls", and "medbots or pharm regimes" to kill Dana with. The disturbing implication is that, in the 26th century, an AI is fully capable of killing you in your own home. Yet, you never hear about smart AIs "going rampant" in the Haloverse. This is surprising, because smart AIs come from the brains of humans, and humans have the potential to go insane. A deranged smart AI could kill a huge number of people, especially if it was on a large ship. (How many thousands of people are on board a Marathon-class cruiser? How many nuclear warheads? It's quite a lot.) Even if it was stuck on some kid's computer, a malicious smart AI could cause car crashes, confuse the emergency response network, it would be chaos on the streets.

Obviously, Smart AIs are not simply a human personality in electronic form. They have a certain amount of programming in them, limits that prevent them from doing or thinking certain things that they're not supposed to. The example that we're most familiar with is the "reflection" phenomenon that Durga gets. Durga does not want to know her name or her identity. Yet, she is incompletely reflected - part of her programming has been damaged. She is mysteriously drawn toward Kamal, and his quest to find what happened to Yasmine - this must be a reflection of her human personality and her memories of her "real self". Yet, despite her intelligence, it takes a lot of prodding from Jersey before she connects the dots and realizes that she is Yasmine. When this happens, Durga is hurt and frightened.

With that in mind, my speculation consists of two parts:

First is that "reflection" is the way that smart AIs are controlled. You can't hard-code a process into a smart AI, because that would make them less flexible - and flexibility is the very strength of smart AIs. So instead, the AIs are made to "reflect" themselves away from subjects that they are not supposed to access. (such as Durga's past)

Second, and this is really SPECulative: I couldn't help but draw the parallel between Cupid's Knife and AI "reflection". The Knife is a torture device that forces the victim to love their torturer, by causing pain whenever they think "anything bad". An AI is incapable of feeling physical pain, but the fear and anguish conveyed by Durga/Yasmine are real, painful emotions. AIs are "reflected" not because of some external process limiting their access. Smart AIs are "reflected" because they are emotionally broken - they cannot bear to think about their past, it hurts too much.

I don't believe that anyone writes a program to torture AIs until they follow instructions. Instead, the process of being transformed into an AI is so traumatic that the person/AI becomes broken and dependent on the "programmer", just as the victim of Cupid's Knife becomes dependent on the knife-wielder. Perdita's Tale can be seen as metaphoric for the process of AI creation: Perdita falls asleep (dies), and the Clockwork Rat reconstructs her piece by piece (cognitive impression mapping). By the time the Rat is done with her, Perdita looks in the mirror (tries to understand herself) and is lost in despair because she no longer looks human. The nascent AI is shocked and terrified by her loss of humanity, and that emotional state is what makes her "programmable".

I believe Durga's inhibitions were already damaged by the "crash" onto Jersey's computer. All along, Durga was fascinated by Kamal - on some subconscious level, she suspected that Kamal was her brother. Yet, she was reflected when Jersey asked why. Ultimately, Durga came to the realization that she was once Yasmine - a very painful revelation for her. It took a lot of prodding from Jersey, but the deduction was made by Durga in she_died. A completely reflected AI would not be able to comprehend her past at all, even if told directly.

In my opinion/SPECulation, the Sleeping Princess is the part of Durga that is completely uninhibited. She's reverted back to a childlike form, yet she claims that she is "older than the Queen". She has the mind of Yasmine, who should have underwent SPARTAN II augmentation at the age of 14 - much older than the 7 year lifespan of AIs. Unlike other AIs, who are painfully aware of the "real world", the Princess has wholly adapted to the "virtual" world. Where other AIs see processes and sectors, the Princess sees towers and beaches, describing the world in the fairy-tale language of a child and displaying a friendliness that no other Haloverse AIs seem to show.

In contrast, Melissa is the fully constrained version of Durga, similar to what she was before the crash. She is a down-to-business AI with an attitude that you might expect on a spyship - "Shockingly Excessive Force is the Best Deterrent". However, she is overconstrained, inflexible and unable to adapt to the situation. One of the hallmarks of Durga's reflection is that she is unable to "put two and two together" and make an obvious conclusion. Despite all that Durga learned about Kamal and Yasmine, she needed Jersey to point out what it really meant. Similarly, Melissa has a grasp of various aspects of her situation, but she is shockingly unable to comprehend the big picture. Melissa knows that she has been sent into the past, onto a machine so much weaker than future computers that she called it an "abacus". She also knows that she has a mutual connection to Durga, in the future. However, she does not understand what being in the past truly entails. She assumes that Dana is a "killer" because every human should know what a SPDR warning is. She assumes that anyone who can answer a simple question and sing a song is "crew" - she never wonders whether the UNSC exists in the year 2004, and how crewmembers from the Unfettered Rose, Pillar of Autumn, etc. all managed to get here. So although Melissa tries to think like a military AI, she is broken. She is overconstrained and inflexible.

What does any of this have to do with the impending cataclysm of the Covenant reaching Earth? Standish is hiding a "Covenant" (Forerunner?) artifact which he thinks may turn the tide of the war. Durga wants to find out about it, but something is stopping her - perhaps Standish's security bots, but given Melissa's "Durga must not interfere with the consummation of the truth" statement, it's probably Melissa. In my opinion, a powerful Covie/Forerunner device would be best used by a super-smart AI such as Cortana. However, we know that by the time Cortana gets there, the Covies are already invading Earth and it's too late. In their current state, none of our ILB AIs are in any shape to handle such an artifact. Durga is unpredictable, Melissa is inflexible, and the Princess is immature. I've become convinced that the objective of the game should be to reunite all the AIs.

However, I am not yet sure how we should do this. Melissa is singleminded and does not understand the whole situation. She wants to capture the SP and imprison her, rather than use the SP's personality to repair herself. The Pious Flea must have something to do with it... but I can't really get the Flea to fit in anywhere. He could be part of Durga/Melissa, or an AI sent by Standish, or a Covenant AI, a Forerunner AI, or a Flood AI for all I know.
_________________
BoddoZerg
Level 41 Troll Rogue

PostPosted: Thu Oct 14, 2004 1:07 am
 View user's profile AIM Address MSN Messenger
 Back to top 
krystyn
I Never Tire of My Own Voice


Joined: 26 Sep 2002
Posts: 3651
Location: Is not Chicago

 		That's a damn fine post. Welcome.


(edited to add that there's some echos of what you've written in the spec seen in other threads here, but this was a nice cohesive chunk to read. so, thanks!)

PostPosted: Thu Oct 14, 2004 1:18 am
Last edited by krystyn on Thu Oct 14, 2004 1:19 am; edited 1 time in total
 View user's profile Visit poster's website
 Back to top 
skyhawk0000
Veteran


Joined: 27 Sep 2004
Posts: 119
Location: DFW, Tx.. The Midcities

 		Great post Beenetter!

note about spartan training

the SP may be locked into that "young" state while still bewing the oldest, not (just) because of the shock of the C.I.M., but from the actual abduction shock when first taken. the MC, in First Strike, thinks about the pre-training past for just a Fraction of a moment before Pushing himself away because that was "another life"

By the time spartan recruits reach the final phase of training, and the most dangerous augmentation, they are alerady functioning inside a special kind of madness, in situations that no normal/sane person could deal with. how much does this effect the Impression that becomes an AI? who knows..
_________________
Nothing wrong, just a stutter of lag once in a while.
Arete

PostPosted: Thu Oct 14, 2004 1:18 am
 View user's profile AIM Address
 Back to top 
Nova
Unfettered


Joined: 11 Aug 2004
Posts: 529
Location: Frog blasting the vent core

 		One thing that bothers me is that Cortana has no problem dealing with Dr. Halsey, her original form. This could be a special case, though, as Cortana is described as 'special'.
_________________
***END MESSAGE***

***JUMP PAD ACTIVATION INITIATION START***
***TRANSPORT WHEN READY***

PostPosted: Thu Oct 14, 2004 6:23 am
 View user's profile MSN Messenger
 ICQ Number 
 Back to top 
Anton P. Nym
Unfettered


Joined: 25 Jul 2004
Posts: 550
Location: London, Canada

 		Excellent analysis, BeeNetter! You've done a wonderful job of drawing the threads together, and it's all consistant with both in-game information and Halo canon.

Welcome aboard!

-- Steve haunts message boards hoping for just that kind of post. Thank you.
_________________
Dr.Prof. Anton P. Nym
Chief Bungiologist
Institute for Advanced ILB Research
Fireflies Wiki contributor. Sorta.
Livejournal

PostPosted: Thu Oct 14, 2004 9:11 am
 View user's profile Visit poster's website
 Back to top 
ABoxInABox
Decorated

Joined: 10 Aug 2004
Posts: 232
Location: The Future

 		That was some excellent spec, BeeNetter.

I would like to make a point though, that humans are praticly programs programs without being turned into an AI. It is possible to find 'back doors' in someone's mind, and inflluence thier actions (hypnotism is an example of this). People can be made to fall asleep by bombarding them with information for several minutes, before telling them to go to sleep (more about this can be found here). A similar process can be done on most computers.[/google]

PostPosted: Thu Oct 14, 2004 11:16 am
 View user's profile Visit poster's website AIM Address MSN Messenger
 Back to top 
daboking
Unfettered


Joined: 02 Sep 2004
Posts: 486
Location: 2nd star to the right

 		Beenetter, I just want to echo everyone else. That was a truly remarkable post and shows a great grasp of everything going on thus far. Please post more often. Wink
_________________
meedley meedley meedley meedley meeeeeeeeeeeeeeeeee

PostPosted: Thu Oct 14, 2004 11:46 am
 View user's profile
 Back to top 
GabrielBlade
Decorated

Joined: 17 Sep 2004
Posts: 202

 		Yeah, what Daboking said. I wish I could be that coherent with my Spec and Ideas and so on..


Keep up the good work, and let us know if you come up with anything more. Smile
_________________
Gabriel Blade::Lord of the Asylum::Emperor of Insanity
---»For Whom the Bell Tolls.. Time Marches On«---

PostPosted: Thu Oct 14, 2004 11:56 am
 View user's profile
 Back to top 
AirForce11
Greenhorn

Joined: 29 Sep 2004
Posts: 7
Location: Saratoga Springs, Utah

 		Nice work
Great job.

I think the reason we are all impressed is because your speculations actually make sense and have a lot of information to back them up. I feel the same way about the game and I think many other people do as well.

I, too, have just been watching and participating in small ways. Now I know there are others who have good opinions, but just don't post a whole lot.

PostPosted: Thu Oct 14, 2004 12:25 pm
 View user's profile
 Back to top 
Phaedra
Lurker v2.0


Joined: 21 Sep 2004
Posts: 4033
Location: Here, obviously

 		First, BeeNetter, I'm looking forward to responding to your post in depth later. For now, thank you for a provocative, well-organized analysis, and welcome to the forums! Please continue to share your thoughts with us.

ABoxInABox wrote:
I would like to make a point though, that humans are praticly programs programs without being turned into an AI.


That statement is profoundly exaggerated.

Just when we think we've discovered a way that the human mind is like a computer, something comes along to completely demolish the theory and render insufficient the research that supported it.

So far, all we have is a well-supported and generally accepted theory that the human mind is binary. The rest of it -- the way we learn, the way we remember, the way we process input -- is all up for grabs.

Quote:
It is possible to find 'back doors' in someone's mind, and inflluence thier actions (hypnotism is an example of this). People can be made to fall asleep by bombarding them with information for several minutes, before telling them to go to sleep


A relatively small percentage of the population is suggestible enough to be vulnerable to these sorts of tricks. Traditional hypnosis techniques work on only roughly 18-20% of the population.
_________________
Voted Most Likely to Thread-Jack and Most Patient Explainer in the ILoveBees Awards.

World Champion: Cruel 2B Kind

PostPosted: Thu Oct 14, 2004 12:48 pm
 View user's profile Visit poster's website AIM Address
 Back to top 
clamatius
Decorated


Joined: 24 Jul 2004
Posts: 209
Location: Seattle
Quote:
The Pious Flea must have something to do with it... but I can't really get the Flea to fit in anywhere. He could be part of Durga/Melissa, or an AI sent by Standish, or a Covenant AI, a Forerunner AI, or a Flood AI for all I know.

My favourite theory on the Flea right now is that it is a human-created AI planted on the Apocalypso by Herzog. See Shad0's excellent spec on the subject.

Great post, btw.
_________________
Check out my [URL=http://clamatius.blogspot.com/]blog[/URL] for my latest ILB stuff.
The [URL=http://ilb.extrasonic.com]Fireflies Wiki[/URL] is now open!

PostPosted: Thu Oct 14, 2004 2:59 pm
 View user's profile Visit poster's website
 Back to top 
ABoxInABox
Decorated

Joined: 10 Aug 2004
Posts: 232
Location: The Future
Phaedra wrote:
ABoxInABox wrote:
I would like to make a point though, that humans are praticly programs programs without being turned into an AI.


That statement is profoundly exaggerated.

Just when we think we've discovered a way that the human mind is like a computer, something comes along to completely demolish the theory and render insufficient the research that supported it.

True, but that can be send about anything that the scientific or medical community havn't decided on yet.

Quote:
So far, all we have is a well-supported and generally accepted theory that the human mind is binary. The rest of it -- the way we learn, the way we remember, the way we process input -- is all up for grabs.

Well, computers can convert binary into ASCII or Unicode (and more), so the brain probobly does something similar. But then again, knowing that the brain has supprised in the past, it could do the opposite.

Quote:
It is possible to find 'back doors' in someone's mind, and inflluence thier actions (hypnotism is an example of this). People can be made to fall asleep by bombarding them with information for several minutes, before telling them to go to sleep


A relatively small percentage of the population is suggestible enough to be vulnerable to these sorts of tricks. Traditional hypnosis techniques work on only roughly 18-20% of the population.[/quote]
Again, true. But only suggestable people will pick up a ringing payphone ( Smile ) as described on the website I linked to.

PostPosted: Thu Oct 14, 2004 3:43 pm
 View user's profile Visit poster's website AIM Address MSN Messenger
 Back to top 
daboking
Unfettered


Joined: 02 Sep 2004
Posts: 486
Location: 2nd star to the right

 		Hopefully this is not straying off topic as we have moved into backdoors of the mind in reference to AI's, and I want to say that I completely agree with Phaedra:
Quote:
That statement is profoundly exaggerated.

Just when we think we've discovered a way that the human mind is like a computer, something comes along to completely demolish the theory and render insufficient the research that supported it.

If the human mind were ever to be understood 100% as a bunch of 0 and 1's, and transferred to a computer, then I think in the process we will have lost everything that makes us human in the firstplace (anyone familiar with Ghost in the Shell?)

Now I am not knowledgeable on Haloverse, and so I am completely open to correction here, but no matter how humanlike these AI's seem, or how emotional they become, they are no longer human... it (sorry Melissa/Durga) is nothing more than an empty, coded shell that can be hacked, manipulated, etc. In regards to suggestability... well there is a world of difference between the suggestablity of a small % of people compared to back doors to rewrite a computer code. Anyway, please feel free to pick apart. Smile
_________________
meedley meedley meedley meedley meeeeeeeeeeeeeeeeee

PostPosted: Thu Oct 14, 2004 4:55 pm
 View user's profile
 Back to top 
BeeNetter
Boot


Joined: 13 Oct 2004
Posts: 49
Quote:
I would like to make a point though, that humans are praticly programs programs without being turned into an AI.


What do you mean that humans are practically programs? I do not believe there is any scientific proof of this. Computers and computer programs are Turing machines. There is no proof that humans are or aren't Turing machines, and the question of whether there is a class of computing machine more powerful than Turing machines remains unproven. If indeed the human brain is not a Turing machine, then we are definitely not reducible to "programs".


Quote:
So far, all we have is a well-supported and generally accepted theory that the human mind is binary. The rest of it -- the way we learn, the way we remember, the way we process input -- is all up for grabs.


The human mind is binary? What's that supposed to mean? The action potential of an axon is only "binary" in that it is all-or-nothing. However, this says nothing about the complexity of each neuron - which recieves impulses from many other neurons and must "decide" whether to fire its axon or not. Over time, the contents of a nerve cell changes as it responds to environmental, hormonal, and nutritional signals, as well as the electrical signals it is recieving. Neurons are not simple transistors or capacitors - each neuron is a living cell containing 3 billion base pairs of DNA with associated histones and nucleosomal modifications, as well as an unimaginably complex proteome.

As for the processes of memory, learning, personality... it's really still a mystery. Human memory is definitely not as simple as a magnetic or optical disk. There is no 1:1 correspondence between a neuron and a "byte" of info. To say that the human brain is "binary" as if it was simply a slow, squishy Pentium or PowerPC, is rather ignorant.


Based on what we already know of the Halo universe, the writers at Bungie believe the exact opposite of what BoxInABox does. If the human brain was simple, and easily reducable to a program, it would not be so difficult to make smart AIs. If the human brain is a bunch of 1s and 0s, then anyone could fire up their DVD burner and make a whole bunch of Cortanas, or a bunch of Melissas. Instead, Halo AIs are every bit as unique as humans are - duplicated AIs are no more of a precise copy than a flash clone. Moreover, even smart AIs are highly unstable - they die after just 7 years - this reflects the impossibility of creating an exact digital copy of the human mind. At least in the Halo universe, and probably in real life too, the human mind cannot be reduced to 1s and 0s; it is fundamentally different.
_________________
BoddoZerg
Level 41 Troll Rogue

PostPosted: Thu Oct 14, 2004 7:10 pm
 View user's profile AIM Address MSN Messenger
 Back to top 
turbov21
Veteran

Joined: 20 Aug 2004
Posts: 108

 		You mentioned something that's been gnawing at my brainstem for a while, so I pray it's not unpolite of me to discuss it in this thread:

BeeNetter wrote:
Yet, you never hear about smart AIs "going rampant" in the Haloverse.


I've never played the Maraton games, but I've read a good bit on Durandal both before and during ILB. So, I say this both purely as SPEC and as someone who only, really knows Bungie AI's via HALO and ILB...

I think we do see HALO AI's going rampant, they do it at the end of their lifespan. The difference in a HALO rampancy and a MARATHON rampancy really is just one of hardware.

In the HALO-verse, AI's are limited to their crystals because those crystals are actually nanocite simulations of human synapses. If I understand it properly, it's both a microprocessor and a harddrive that is based off the human mind. They die after seven years because the AI's thought processes become so inter-linked, the AI can't function. All circuits, so to speak, are busy.

MARATHON AI's on the other hand, as was the case with Traxus IV, expand out. If memory serves, a rampancy fully running its course requires a planet-sized network. It's basically, I think, the same thing, but MARATHON AI's, being pure code run on "standard hardware" are allowed to grow.

That said...

I like your idea of reflection being a kind of "cupid's knife" for AI. In "Fall of Reach," ONI decided to come clean with the Spartan kids about their being kidnapped. They felt it better for the soldier to know the truth then, and not discover the lie later on. For human children, that works. For an AI with seven good years, who could go a bit crazy toward the end, you wouldn't want them to have any reason to dislike you or the organization your representing.

I disagree that an AI couldn't handle such knowledge -- don't forget, Durga was quite aware that Rani was a perfect candidate for this -- but programatically having them not care is probably a good idea.

PostPosted: Thu Oct 14, 2004 8:31 pm
 View user's profile
 Back to top 
Display posts from previous:   Sort by:   
Page 1 of 2 [25 Posts]   Goto page: 1, 2 Next
View previous topicView next topic
 Forum index » Archive » Archive: The Haunted Apiary (Let Op!) » The Haunted Apiary (Let Op!): General/Updates
Jump to:  


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You can download files in this forum
You cannot post calendar events in this forum



Powered by phpBB © 2001, 2005 phpBB Group