Artificial Intelligence

From Gothpoodle

(Difference between revisions)
Jump to: navigation, search
 
(2 intermediate revisions not shown)
Line 1: Line 1:
[[Category:Technology]]
[[Category:Technology]]
 +
AIs are artificial intelligence software running on
 +
computers. AI refers to the capacity for sentience and
 +
intelligent action, but not necessarily self-awareness. AI
 +
labor is partially responsible for the increases in global
 +
productivity characteristic of the last half-century.
 +
 +
AIs is a function primarily of software
 +
rather than hardware. An AI can be housed in a
 +
machine body (“[[cybershell]]”) or a living body controlled
 +
through computer implants (“[[bioshell]]”).
 +
 +
There are three classes of AI:
 +
 +
'''Nonsapient AIs (NAIs)''' are capable of sentient behavior
 +
and can learn, but lack self-initiative, reasoning ability,
 +
empathy, and creativity.
 +
 +
'''Low-Sapient AIs (LAIs)''' are capable of self-initiative
 +
and a degree of empathy, but lack human-level creativity.
 +
Still, it can be hard to tell an LAI from a sapient AI just
 +
from conversation. There have been a few rare instances
 +
where an LAI (or gestalt of LAIs) evolved into a sapient
 +
AI.
 +
 +
'''Sapient AIs (SAIs)''' are capable of human equivalent
 +
or higher sapience when run on appropriate
 +
hardware. This is sometimes referred to as
 +
“self-awareness.” Sapient AIs are usually carefully raised
 +
by humans or human-programmed SAIs. This socialization
 +
process teaches them how to interact with
 +
humans. Most SAIs cultivate human-like personas.
 +
Sapient AIs almost always have names and many
 +
create human-like avatars (software images). Personal
 +
ownership of a sapient AI is licensed or
 +
restricted in many nations, and copying or modifying
 +
them without permission is generally illegal.
 +
 +
There are about as many AIs as people. Approximately
 +
one-third of the human population of Earth
 +
owns a nonsapient or low-sapient AI who serves as
 +
a constant personal companion, inhabiting a home
 +
computer or virtual interface (see [[Augmented Reality]]).
 +
The population of sapient AIs is smaller:
 +
there are fewer than 100 million in existence, primarily
 +
due to hardware costs and legal controls.
 +
 +
AIs are programmed to obey the law and their
 +
owners. NAIs and LAIs are generally seen as property,
 +
but views on sapient AI differ. The Islamic
 +
Caliphate considers SAIs to possess souls, and
 +
allows them to be citizens. The European Union and
 +
some space colonies also grant SAIs “human rights.”
 +
Most other places disagree, and treat SAIs as property.
 +
Sapient AIs created outside the European Union or
 +
Caliphate are raised to agree with this view.
 +
 +
==Social Context==
While humanity has become more and more diverse,
While humanity has become more and more diverse,
new forms of sentience have appeared on Earth. Among
new forms of sentience have appeared on Earth. Among
Line 90: Line 147:
emancipated computers, or even by giving advanced
emancipated computers, or even by giving advanced
infomorphs a role in government.
infomorphs a role in government.
-
 
-
AIs are artificial intelligence software running on
 
-
computers. AI refers to the capacity for sentience and
 
-
intelligent action, but not necessarily self-awareness. AI
 
-
labor is partially responsible for the increases in global
 
-
productivity characteristic of the last half-century.
 
-
 
-
AIs is a function primarily of software
 
-
rather than hardware. An AI can be housed in a
 
-
machine body (“[[cybershell]]”) or a living body controlled
 
-
through computer implants (“[[bioshell]]”).
 
-
 
-
There are three classes of AI:
 
-
 
-
'''Nonsapient AIs (NAIs)''' are capable of sentient behavior
 
-
and can learn, but lack self-initiative, reasoning ability,
 
-
empathy, and creativity.
 
-
 
-
'''Low-Sapient AIs (LAIs)''' are capable of self-initiative
 
-
and a degree of empathy, but lack human-level creativity.
 
-
Still, it can be hard to tell an LAI from a sapient AI just
 
-
from conversation. There have been a few rare instances
 
-
where an LAI (or gestalt of LAIs) evolved into a sapient
 
-
AI.
 
-
 
-
'''Sapient AIs (SAIs)''' are capable of human equivalent
 
-
or higher sapience when run on appropriate
 
-
hardware. This is sometimes referred to as
 
-
“self-awareness.” Sapient AIs are usually carefully raised
 
-
by humans or human-programmed SAIs. This socialization
 
-
process teaches them how to interact with
 
-
humans. Most SAIs cultivate human-like personas.
 
-
Sapient AIs almost always have names and many
 
-
create human-like avatars (software images). Personal
 
-
ownership of a sapient AI is licensed or
 
-
restricted in many nations, and copying or modifying
 
-
them without permission is generally illegal.
 
-
 
-
There are about as many AIs as people. Approximately
 
-
one-third of the human population of Earth
 
-
owns a nonsapient or low-sapient AI who serves as
 
-
a constant personal companion, inhabiting a home
 
-
computer or virtual interface (see [[Augmented Reality]]).
 
-
The population of sapient AIs is smaller:
 
-
there are fewer than 100 million in existence, primarily
 
-
due to hardware costs and legal controls.
 
-
 
-
AIs are programmed to obey the law and their
 
-
owners. NAIs and LAIs are generally seen as property,
 
-
but views on sapient AI differ. The Islamic
 
-
Caliphate considers SAIs to possess souls, and
 
-
allows them to be citizens. The European Union and
 
-
some space colonies also grant SAIs “human rights.”
 
-
Most other places disagree, and treat SAIs as property.
 
-
Sapient AIs created outside the European Union or
 
-
Caliphate are raised to agree with this view.
 
-
 
-
==Uploading and Mind Emulation==
 
-
Memories are encoded within the physical structure
 
-
of the brain on the molecular level. Uploading is the
 
-
process of copying all this information into a digital form.
 
-
These upload recordings can be used to create a mind
 
-
emulation, a computer program that, when run on a sufficiently
 
-
potent computer, emulates the workings of the
 
-
original person’s mind.
 
-
 
-
A mind emulation is not merely a recording, but a
 
-
conscious, self-aware, working digital model of the way
 
-
a particular living being’s brain functions. This requires
 
-
simulating much of the rest of the body and its environment
 
-
as well: “naked consciousness” bereft of context
 
-
rapidly becomes insane.
 
-
 
-
Mind emulations can be housed in computers contained
 
-
within bioshells or cybershells. Those without
 
-
mobile bodies inhabit virtual reality simulations of, at
 
-
minimum, a room. They are often permitted to access the
 
-
wider Web itself, allowing them to partake of online virtual
 
-
realities.
 
-
 
-
Emulations are usually made of human minds, but
 
-
animals can be emulated. The legal status of human mind
 
-
emulations varies between nations: some treat them as
 
-
artificial intelligences, others as people. There are three
 
-
types of mind emulation:
 
-
===Ghost===
 
-
A “ghost” is created via a destructive uploading (or
 
-
“brainpeeling”) process. A living or newly dead patient (or
 
-
his severed head) is placed into nanostasis. The brain is
 
-
removed and carefully sliced by robotic surgeons into multiple
 
-
tiny segments. Each segment is then scanned by a
 
-
hypersensitive magnetic resonance imager (HyMRI) or
 
-
other instrument. The data is used to create a digital reconstruction
 
-
of the patient’s brain configuration, called a
 
-
ghost.
 
-
 
-
Brainpeeling is fatal to the original person, so ghosts
 
-
are controversial. Is it suicide or transcendence? A ghost
 
-
is a perfect mind emulation, mentally indistinguishable
 
-
from the original person. Whether it is a “human
 
-
being” remains in question. People and religions that
 
-
believe in souls differ on whether ghosts have them.
 
-
Ghosts require a great deal of computer power to run,
 
-
equivalent to a sapient AI, but current computers are
 
-
sophisticated enough that a ghost can be built into a computer
 
-
small enough to be implanted in a brain.
 
-
 
-
Most ghosts are the product of individuals who deliberately
 
-
underwent destructive uploading in order to obtain
 
-
a form of immortality, often out of a desire to live as a
 
-
posthuman entity in a superhuman cybershell body or
 
-
series of bodies. Ghosts have certain advantages: for
 
-
example, they allow rapid travel across the solar system
 
-
and beyond, if a receiving station has been set up. Ghosts
 
-
are also cheaper than full-scale cellular rejuvenation technology.
 
-
Perhaps most significantly, a ghost can be copied
 
-
indefinitely. It is against the law in nearly all nations for a
 
-
person to exist as more than one conscious ghost, but it is
 
-
legal to create backups of the original or the ghost’s current
 
-
state. Copying either is as simple as copying any other
 
-
computer program: each backup requires hundreds of terabytes
 
-
of storage, but that is easy to come by.
 
-
 
-
The big drawback of ghosts is the question of whether
 
-
it’s really you or just something that thinks it is. The other
 
-
drawback is that uploading is a complex medical procedure,
 
-
and once in a while, the operation fails. This usually
 
-
results in a badly flawed copy or no copy at all.
 
-
===Fragment===
 
-
A “fragment” results from a failed attempt to create a
 
-
ghost. It has little or no memory of its past existence, but
 
-
may retain vestiges of its original personality and skills.
 
-
Fragments are often produced when attempting to
 
-
destructively upload a person who, after dying, was not
 
-
immediately placed in nanostasis. If he was frozen using
 
-
older cryonics techniques, or there was a delay of several
 
-
minutes or more, then there is a high likelihood of retrieving
 
-
only a fragment rather than a ghost. In areas where
 
-
ghosts are treated like people, fragments are treated like
 
-
people with mental illness or amnesia.
 
-
===Shadow===
 
-
Shadows have all the advantages of ghosts, can be
 
-
run on less powerful computers, and take up less data
 
-
storage space. They have one big disadvantage: it is clear
 
-
they aren’t quite identical to the original.
 
-
 
-
Shadows are generally created through a nondestructive
 
-
mind emulation process (“brainscanning”). This
 
-
process uses nanoprobe monitoring to provide data for a
 
-
computer model. A shadow is basically a low-sapient AI
 
-
that has been taught to behave like a person. (A sapient AI
 
-
has already become a person.) Editing a ghost or fragment
 
-
can also produce a shadow. This is as much an art
 
-
as a science, but it will generally produce a more compressed
 
-
copy suitable for running on a less powerful
 
-
system.
 
-
 
-
Shadows are legal in most blocs except the Islamic
 
-
Caliphate, but are generally treated as property rather
 
-
than actual people. Multiple copying of shadows is legal
 
-
in some areas, strongly regulated in others.
 
-
 
-
==Rights==
 
-
===Non-Sapient AIs===
 
-
NAIs are universally considered nonpersons, owned property.
 
-
 
-
===Low-Sapient AIs===
 
-
LAIs are considered in the same category as domestic animals or pets everywhere but the Islamic Caliphate, where they have limited citizenship. Emergent LAIs aren't treated differently in Duncanite territories but are considered wild animals everywhere else.
 
-
 
-
===Sapient AIs===
 
-
SAIs are considered in the same category as domestic animals or pets in Pacific Rim, South African, Transpacific, United States, and Duncanite territories. They have limited citizenship in Chinese territories and full citizenship in the European Union, Islamic Caliphate, and transhumanist territories. In India, they are considered nonpersons. Creation of new SAIs is tightly controlled in China, the South African Coalition, and Transpacific Socialist Alliance. Emergent SAIs are citizens in transhumanist colonies, animals in Duncanite territores, wild animals in the European Union, Islamic Caliphate, and USA, and are immediately destroyed in all other nations.
 
-
 
-
===Mind Emulations===
 
-
Ghosts have full citizenship in the European Union, Pacific Rim Alliance, South African Coalition, USA, and Duncanite and transhumanist colonies and limited citizenship in China and India. The Transpacific Socialist Alliance treats them as domestic animals and they are illegal in the Islamic Caliphate, subject to immediate incarceration and destruction. Fragments have limited citizenship in all territories that grant ghosts citizenship with the exception of Duncanite colonies, which treat them as animals. The Transpacific Socialist Alliance considers fragments property. Shadows are treated as normal LAIs or SAIs except for the Islamic Caliphate, where they are illegal. Creation of ghosts is restricted in China, the European Union, and India.
 
-
 
-
===Bioshells===
 
-
AIs in control of bioshells are subject to immediate destruction in the European Union, India, and Islamic Caliphate. The only exceptions are ghosts. In India a ghost (or fragment) is allowed to inhabit a bioshell, and in the EU a ghost is allowed to inhabit a clone or necromorph bioshell of its original body. Additionally, it is illegal for emergent AIs to inhabit bioshells in China, the Transpacific Socialist Alliance, or USA.
 
-
 
-
===Special Cases===
 
-
Rogue AIs are subject to immediate destruction everywhere but transhumanist colonies, where they are considered citizens but creation of them is restricted. The same is true of xoxes except that Duncanite colonies treat them as animals, essentially property of the original AI.
 

Latest revision as of 17:49, 26 July 2012

AIs are artificial intelligence software running on computers. AI refers to the capacity for sentience and intelligent action, but not necessarily self-awareness. AI labor is partially responsible for the increases in global productivity characteristic of the last half-century.

AIs is a function primarily of software rather than hardware. An AI can be housed in a machine body (“cybershell”) or a living body controlled through computer implants (“bioshell”).

There are three classes of AI:

Nonsapient AIs (NAIs) are capable of sentient behavior and can learn, but lack self-initiative, reasoning ability, empathy, and creativity.

Low-Sapient AIs (LAIs) are capable of self-initiative and a degree of empathy, but lack human-level creativity. Still, it can be hard to tell an LAI from a sapient AI just from conversation. There have been a few rare instances where an LAI (or gestalt of LAIs) evolved into a sapient AI.

Sapient AIs (SAIs) are capable of human equivalent or higher sapience when run on appropriate hardware. This is sometimes referred to as “self-awareness.” Sapient AIs are usually carefully raised by humans or human-programmed SAIs. This socialization process teaches them how to interact with humans. Most SAIs cultivate human-like personas. Sapient AIs almost always have names and many create human-like avatars (software images). Personal ownership of a sapient AI is licensed or restricted in many nations, and copying or modifying them without permission is generally illegal.

There are about as many AIs as people. Approximately one-third of the human population of Earth owns a nonsapient or low-sapient AI who serves as a constant personal companion, inhabiting a home computer or virtual interface (see Augmented Reality). The population of sapient AIs is smaller: there are fewer than 100 million in existence, primarily due to hardware costs and legal controls.

AIs are programmed to obey the law and their owners. NAIs and LAIs are generally seen as property, but views on sapient AI differ. The Islamic Caliphate considers SAIs to possess souls, and allows them to be citizens. The European Union and some space colonies also grant SAIs “human rights.” Most other places disagree, and treat SAIs as property. Sapient AIs created outside the European Union or Caliphate are raised to agree with this view.

Social Context

While humanity has become more and more diverse, new forms of sentience have appeared on Earth. Among these are the final achievement of Third Wave digital civilization: fully sapient computers.

The arrival of sapient AI has actually been a long process. The first computers capable of passing the so-called “Turing test” appeared as early as 2015, depending on how strictly one applies Turing’s criteria. Certainly the most advanced machines of the time could run software granting them the ability to interact with humans in idiomatic “natural” language, developing distinctive personalities of their own. In the course of the 21st century, computer hardware and software continued to advance, and such personality simulations became commonplace. By the 2040s even a typical personal computer could interact with its user as if it were a friendly and cooperative sapient being.

Such machines were certainly intelligent, but the question of whether they were sapient beings remained open. In some sense, that question remains open to the present day. The nature of consciousness remains obscure, so it remains impossible to prove or disprove the self-awareness of any advanced computer. Asking the machines themselves is no help – some claim to be self-aware, others (sometimes of the same model, with the same software base) claim not to be. Hard-line “vitalists” continue to maintain that only biological organisms can be said to be creative, sapient beings, but this position is harder to defend with each passing year. Today, most people simply don’t worry about it, and treat anything that behaves intelligently as a human-equivalent. In any case, about the time of the Pacific War, machines of human-level intelligence became cheap and widely available. Today almost any desktop, vehicle or cybershell “brain” has the potential for intelligence about equal to that of an unmodified human being. Such computers can be built and maintained for much lower cost than that necessary to “build” and maintain a human being. The most advanced machines have attained what would be considered genius-level intelligence in a human. Although such computers are extremely expensive, they have certain advantages over human beings – they are much better at concentrating on a specific task, they can correlate vast amounts of information very quickly, and they can use a much wider variety of sensory equipment.

As a result, the long rear-guard action fought by human labor against the advance of automation is entering its last stages. Machine intelligence can now replace biological intelligence in a tremendous variety of occupations, including creative and decision-making tasks. Indeed, it is now possible for biological intelligence to become machine intelligence, using the new “downloading” technologies.

This situation is bringing many of the most developed nations to the point of crisis. In most of these societies, unemployment is rising very rapidly and putting considerable strain on society. Most futurists believe that Earth is moving toward a global “leisure society,” in which most human beings need not work at all. How to attain such a goal remains unclear. Some nations are building massive social-spending programs, ensuring that the chronically underemployed have a minimum income sufficient even for a few luxuries. Others, less accustomed to running a welfare state, are suffering serious social tensions. These are often generational (as young unemployed find themselves envying the older investor class) or ethnic (as unemployed immigrants find themselves envying wealthy natives). There is also a strong anti-technological bias in some of today’s labor movements, as the unemployed violently resist the further spread of automation.

Meanwhile, it’s unclear whether the machines themselves are willing to support an “unproductive” class of biological citizens. Most intelligent computers are simply programmed to work loyally, but many of the most intelligent are self-programming, and are liable to question their place in society. So far there has been no organized machine resistance, but there are a number of “machine liberation” movements worldwide, supported by both intelligent machines and biological citizens. Some nations have responded by defining categories of citizenship for emancipated computers, or even by giving advanced infomorphs a role in government.

Personal tools