Virtual Intelligence

WORK IN PROGRESS

Virtual Intelligence, better known by the acronym "VI", is a form of neuromorphic computer meant to mimic real life. VI mimic how organic creatures process information, rather than operating like traditional computers. Many people erroneously use the term "artificial intelligence" or "AI" to describe these machines, but that is incorrect. While they closely mimic a real brain, they are not sapient and are, just like traditional computers, slaves to their own programming's limits, even if they are capable of learning and improving themselves to a point. The philosophy of "sapience" itself changed drastically by the invention of VIs, as did "sentience". More on that later.

Neuromorphic Processing

A traditional computer has discrete components operating in tandem in order to perform computations. For instance, it has a central processing unit to perform calculations, a memory module to hold data that needs to be processed, long term memory to store long-term data. The key bottleneck is the transfer of information between these different components. Several ways have been implemented to get around this, such as local cache memory, but all that does is extend the time before the bottleneck is hit. It's also extraordinarily energy inefficient, requiring large amounts of power to operate, leading to a lot of heat.

Mimicking the brain

An organic brain does not operate in that manner. It has millions of individual neurons instead of discrete components. There is no central memory storage, no long term storage, no central processing unit. Instead, it operates on a neural state system, where each neuron operates in interconnected states. A traditional transistor in a classic computer has no memory of what charge has been sent through it, hence a separate memory module is needed. A neuron functions as a "memristor". Basically a transistor with a memory of what current passed through. This allowed the design of a synthetic neuron, and thus making artifical synapses.

The artifical brain

A neuromorphic computer still operates out of programming and a set of parameters. It is no more alive or sapient than a traditional computer is. However, the way it processes information is ridiculously efficient and fast in comparison and its capacity for self-improvement and learning is vastly superior to a traditional computer.

Practical application

With the advent of the neuromorphic computer, VI became a possibility. It possessed the capability to process and handle information on a "subconscious", "effortless" or more correctly "energy efficient" level, allowing it to mimic living organisms to a near perfect degree. A properly configured and trained VI could easily pass as a living being in a conversation, provided the living person interacting with it could not see something that immediately gave it away. Casual onversation, even with live voice synthesis, is near perfect with high grade VIs.

This opens up several venues for faster and more efficient processing of certain types of information, making VI's exceptionally skilled in niche areas of computing and operation. Still, traditional computers are still far more prevalent, cheaper to construct and far easier to program for.

Inventor
Unknown (information lost in The Cataclysm)
Claimant and patent holder
Seris Incorporated
Discovery
2612 AC

Related Reading

Ly-R-AIn-E
Character | Jan 19, 2021

Ly-R-AIn-E is an advanced virtual intelligence that has been run inside the Grand Archives since long before the Cataclysm. With her extensive knowledge and long run-time, she behaves almost as if she was sapient.

Combat Robotics
Technology / Science | Apr 12, 2021

Combat Robots are semi-autonomous droids designed to fight or provide security. Often durable and with lightning fast reflexes, combat droids are a real threat on the battlefield.

Needing to learn

The drawback of VIs is that neuromorphic computing requires a lot of expensive materials and competence to design them. In addition, VIs can get a base programming, but because they are created in a starter state, they aren't very capable when first booted. They need to operate and learn in order to establish the neural pathways required to perform their designated function, much like children need to learn. It's obviously much faster to teach a VI and let it learn its function than it is to teach a child, but it still takes time and you can't just place a VI in a function and expect it to perform flawlessly immediately. Initially, a VI performs worse than a traditional computer, until it reaches a point where its neural pathways and synapses have been established, allowing it to use the same shortcuts living brains do to efficiently process data.

Mistakes

Because of its design, a VI can learn bad behaviours and is not deterministic. Generally, they are very reliable, but when exact numbers and maths is required, a traditional computer is more reliable. The neuromorphic VI is usually many times faster and can take into account linguistics, tone and all the information stored in its neural system, but the traditional computer can always be relied on to perform its programming to a T, whereas a VI might make mistakes because of their complexity.

For this reason, in combination with their extreme cost, traditional computers are still dominant in the world and are used everywhere where a VI isn't a clearly superior choice.

Known virtual intelligences

Since VIs tend to develop personalities and accrue experience, they are also often assigned an identity by its operators. One of the world's most known VIs is Ly-R-AIn-E, a web-based search VI operated by Seris Electronics. LyR is known to be snarky, elitist and lacking patience, but it is unmatched in its efficiency in finding information and interpreting the needs of the users. LyR is more than a hundred years in operation, and in spite of damage to its synapse network in the Cataclysm, LyR is one of the prime examples of what a VI can do that traditional computers cannot.

Combat Robots

Combat robots, better known as droids, is another area where VIs are used extensively. Most models from Marek Accoutrement have the capability to mount an on-board VI, or operate via synclink to a centralized VI. The VI then operates in a sort of distributed hivemind, or rather a central mind with many bodies. The VI must be trained to operate in this mode, as handling that many sensory and motoric inputs is something it's not well versed to do initially.

Most droids will operate with a VI only when it's necessary, such as when they operate in Hunter-Killer mode or when the operator determines that the distributed VI mode is necessary. Most operators have a VI trained to operate in this mode and it can be connected at any point. Usually, droids operate with an on-board computer that handles communication, control of the robotic body and sensory input, receiving inputs from a centralized computer. That computer is usually a traditional supercomputer operating from a set list of parameters. The VIs are only used in difficult situations where creativity and responsiveness is required.

On the topic of sapience

While virtual intelligences are undoubtedly classified as "intelligent agents", whether or not they can be considered sentient or sapient is a subject of much debate. Sentience, meaning the ability to "feel" something, doesn't just concern itself with feelings of joy, happiness, indifference, love or anger for instance, but all manner of feeling. This includes the ability to feel touch, smell something or see something.

In that sense, VI's are undoubtedly sentient, as they are more than capable of sensing their surroundings using various sensors. They also understand feelings and emotions to a greater degree than most of the world's population, being able to mimic them accurately based on situation.

Does that mean that VI's actually feel something? Or are they mimicking a learned behaviour? Some argue that the question is moot, as organic beings are controlled by an internal regulation system based on nerve signals and hormones. They are, in effect, programmed, to "feel" certain emotions at certain times, which makes them behave in a particular manner.

Are programmed VI's any different? They have a set of internal instructions that react according to external stimuli. Whether or not the VI does this "consciously" (by choice) or automatically (by programming) is largely irrelevant. The effect is that they can experience and express emotion. They are also fully able to suppress emotion, much like organics do.

The conclusion, for most philosophers, is that VI's undoubtedly can be considered intelligent and sentient.

About sapience

That leaves the question, of course, if VI's are sapient. Sapience, like sentience, has many different definitions and "grades", with the organic existance of being self-aware, as a Kykr, is considered the highest form of sapience.

VI's are not recognized, neither by definition nor by law, as sapient. They are tools, not people. Machines, not individuals. While individiual VI's develop personalities and appear to be unique from one another, a VI with the same programming, put through the same learning experiences, will develop most exactly the same.

Can't the same be said for people? Absolutely. By this measure, VI's are sapient. What defines the VI's non-sapiency is the fact that while they may be both sentient and self-aware, they have no agency. No passion or drive of their own. In addition, they are completely void of self-determination.

VI's are given a playfield of where they can make decisions for themselves, in line with their programming. A person's playfield changes over time. Their opinions and morals shift based on both experience and external stimuli.

A VI is incapable of that form of development. A VI can never directly deviate from their programming. One very real example of this is Ly-R-AIn-e, who, in spite of her outwardly disdain for anything organic, has never ever refused a single command from said organics. She is sentient, she feels disdain. But she is not sapient. She has no ability, will or even desire to disobey her programming.

In essence, most philosophers, and engineers, consider VI's sentient and partially sapient, but cannot be considered real sapients, worthy of individual recognition and rights. While organics are partially slaves to their own programming (evolution, hormones, neural signals, learned behaviours etc), VI's are fully slaved to their programming. They can learn and improve, but never evolve or decide for themselves.

Given freedom and individual rights, they wouldn't know what to do with themselves. They do not understand the concept.


Articles under Virtual Intelligence



Cover image: by CC BY, Modified by Tobias Linder

Comments

Please Login in order to comment!
20 May, 2021 14:43

I love the idea tha virtual intelligence systems can acquire personalities given enough time. :D

Powered by World Anvil