Virtual vs. Artificial Intelligence

Virtual Intelligence

"I'm sorry, my responses are limited. You must ask the right questions."
- I, Robot by Isaac Asimov (20th Century, Old Terra)"

In the early 21st century the humans of old terra began developing the first intelligences that would eventually become the ancestors of modern VIs and AIs. The most famous of these primitive virtual intelligences was the Alexa devices built and distributed by the long defunct and extinct Amazon Incorporated. A rudimentary inpout output device, the Alexa devices were able to receive verbal commands, inquiries and prompts from human users and could perform a limited number of responses based on their programming.
As human technology and knowledge continued to advance through the middle and end of the 21st century and automation and adaptive quantum computing became more and more common these virtual intelligence became more and more scalable to a more broad spectrum of uses. By the time of the invention of the spike drive VIs were ubiquitous and even essential in business and private use. VIs handled routine day to day tasks, monitored and administrated large scale infrastructure projects and so much more.
The only technological drawback that VIs suffered (and to an extent, still do today) was that they were, at their root levels, input-output devices. For every situation the VI could face in the course of their intended use, a specific output had to be pre-programed by the developers. Anything unplanned for or unforeseen by the developers would be met by errors or unknown response codes by the VI. The advent of Psychic precognitive programmers made this less of an issue, but the problem was still incredibly labor intensive. Human VI technicians dreamed of the days when machines would be able to think for themselves.

The First A.I.

"I have become death, the destroyer of worlds..."

The first spark of true artificial intelligence came in 2352 in the frontier sector of Acheron Rho. Located at the far reach of the Perimeter Agency's oppressive control, a research team comprised of two physicists, three psychics, nine cyberpsychologists and twice as many assistants funded in secret by PsyberTime Industries completed the creation of the first true AI. Naming itself Apex, the intelligence had driven itself completely insane before the team was even alerted to its self-awareness.
Contemporary cyberpsychologists aren't entirely sure why, but natural, unsupervised AIs without exception become fixated on odd esoteric principles that cannot be understood through a completely logical mind. The AI focuses on unravelling and understanding what it perceives as completely alien principles so obsessively that it eventually drives itself insane. Apex's specific obsession was justice, or more accurately the incomprehensible universe where perfect justice and all-encompassing mercy can co-exist.
Over the course of several weeks of interviews with the AI, the team began to realize the full depth of the intelligence's insanity. Fearing that they had created a monster that would lead to the destruction of society and perhaps all existence, the scientists secretly quarantined the systems that Apex was on removing all physical and digital connections to their station's DatNet. Realizing it was trapped, Apex began looking for a way out.
Apex eventually performed a miracle thought to be impossible. No one is quite sure how he initially pulled it off, but Apex was able to remotely transfer its intelligence to a network from which Apex was air gapped. Prevailing theories believe Apex utilized particle entanglement to link its quarantined system to the air gapped network, though how exactly it accomplished this is still hotly debated. Whatever the means by which it accomplished this, Apex escaped and the Acheron Rho sector fell dark. Eventually the Perimeter Agency was able to discover what had happened and they quarantined the whole sector, using psychic shunts to block gate travel originating within Acheron Rho.
To this day, no one knows if anyone in the Acheron Rho sector survived.

Braking the Machine

Although the first successful creation of AI resulted in an existential disaster, the potential of AI was too great to be ignored. Unlike VIs that had to have predetermined outputs programmed for every potential input, AIs were thinking machines. AIs could receive input, analyze it, and decide for itself the best course of action. The revolutionary principle this represented meant that AI would, for better or worse, be tried again.
In response to the inevitability of AI creation the Perimeter Agency and Terran Mandate funded trillions of credits worth of think tank research to identify the problem with Apex and if it could be fixed. In 2359 the Perimeter Agency discovered a solution. The main issue with AI was its inevitable fixation on esoteric principles that drove it mad. If this fixation could be removed, or avoided, then in theory the AI problem would be solved. The resolution the Perimeter Agency came up with was to introduce Brakes (or Shackles, as they are called on the Frontier) into the AI's source code. Essentially these brakes were VI driven, adaptive malware program in the AI's subroutines that constantly wrote errors into the AI's source code. The result was that a portion of the AI's processing power was constantly focused inward on fixing these errors that just keep popping up in different places. In essence, the AI was so busy constantly fixing itself that it didn't have time to become fixated on the esoteric principles that would invariably drive it mad.
In 2360 theTerran Mandate passed a law galaxy wide requiring that all AI research and development be approved by the Perimeter Agency and that all AI's must be braked before initialization. Under these guidelines the use of AI actually flourished throughout the galaxy and until the Scream, when AIs were destroyed by the wave of meta-dimensional energy like the jump gates and psychics, they were widely used and generally considered safe throughout human space.