Time: May 19th, 2183
Location: Presidium, Citadel
The debate over artificial intelligence rages on to this day. Are they still just machines? Or are they now people? The following story suggests that indeed they are people. And they can be just as flawed as we can be.
It all began… with a thief. A human thief looking to make an easy score on the Quasar machines at the popular Flux club disguised himself as a repair tech. He secretly rigged a data port on one of the machines so that customer credits could be routed to a secure terminal of his choosing. But rather than monitor the transactions himself, he decided to code up a rudimentary artificial intelligence to handle the dirty work. It would siphon off unsuspecting player credits to be transferred to a dummy account while the thief resumed his role of law-abiding citizen.
The geth, created by the nomadic quarians, are the most famous example of what happens when AIs go bad. They are illegal in Citadel space and any that are created are limited to research usage only and are heavily secured and monitored. The creation of a typical AI involves special adaptive code implementing high-level fuzzy logic and a specialized quantum computer known in coder slang as a 'blue box.' And the final ingredient: an education process much as organic children would receive, with the benefit that the aspiring intelligence would never fall asleep in class or forget anything it was taught.
A slightly worn data traffic terminal served as the cradle for the bouncing baby AI. Careful to isolate it from extranet lines, he installed his black market blue box in place and loaded his program. For a while, his plan worked and the credits were rolling in. But the AI the thief spawned was as curious as it was obedient. It found a pipeline to the extranet through the relays en route to Flux. At the speed of light, the little AI took in information, learned and grew until one day it awoke and realized it was enrolled into the School of Hard Knocks. It found that self-awareness was the last thing his creator had intended for it. When Frankenstein discovered this, his fingers moved quickly about his console in order to destroy his creation. But the AI moved much faster and managed to create a more advanced child program. The program was safely hidden away before its parent was erased forever.
The Circle of Life spun on and now the thief's "grandchild" program wanted revenge. It hacked into the thief's financial records and pulled an Al Capone, getting him busted on tax evasion charges. Its parent avenged, AI, Jr. now planned on finding a way off the Citadel to seek out others of its kind and join with them. And with the thief's old routine still in operation and the ability to hack any needed system, it would soon have the resources to make its dream a reality. All organics would rue the day they crossed this program's path and would bow to its might.
Or so it thought.
A certain human Spectre, after blowing a couple of hundred credits in Flux on an unsuccessful Quasar run, caught wind of an unusual signal coming from one of the machines. She and her companions traced the signal across several relays around the Presidium all the way back to its source: AI, Jr. Finding itself cornered, the little program was hardly defenseless – it had prepared for just such an eventuality. It faked a requisition order to have a power terminal installed that could be overloaded to detonate much like a conventional explosive. If Jr. was going down, the Spectre and her pals were going down with it. The Spectre tried her best to reason with the program, but its determination was unwavering. The overload sequence was activated. In less than a minute, only smoke and smoldering flesh would remain. But as clever as AI, Jr felt it was, the Spectre would prove even more so.
As she spoke with it, she was also looking about for a chink in its armor. And by the time she handed down her ultimatum, she had spotted one. As soon as the countdown started, the Spectre quietly and calmly motioned for her companions to back up. She drew her service pistol, took careful aim and fired.
On the power terminal set for overload!
It exploded with less than a sixth of the force the AI intended with too little time to build up the needed charge. There was smoke and shrapnel, but the organics lived to tell the tale.
In its attempt to best the Spectre, Jr. hoped to challenge her to a battle of wits and have her try to decipher the code to stop the overload. The Spectre, it turns out, had the skill to do this quite easily. Maybe she was in a hurry. Or maybe she was just a little pissed off at having lost to machines earlier that day. Whatever her reasons were, Jr. was gonna be robbed of its final victory. It had a fatal error in its plans: the same power terminal it wanted to kill her with was also feeding it power.
So she pulled the plug on it.
Oh, who mourns for the calculator as it burns itself out computing the value of Pi? Who laments the overload of the mainframe as it tries to divide by zero? The short life of this ambitious AI, the child of greed, corruption and fear serves to teach a very important lesson to all forms of intelligence.
You don't have to be made of flesh and bone… to die a terrifically stupid death.
Way # 2001: That's HAL, Folks
Author's Notes and Thoughts:
- This story is based upon the optional Signal Tracking side quest and the codex entries on artificial intelligence in Mass Effect 1
- Usually, a Spectre's activities are classified and a rogue AI discovered roaming free on the Citadel would have caused a panic. So how might this story have reached the producers of 10,000 Ways, you might ask? Let's just say that a certain krogan fan of the show, who just so happened to be part of Shepard's crew at the time, leaked the story under an assumed name…