Why Don't We Have Artificial Intelligence Yet?

QUESTION: Why don't we have Artificial Intelligence yet?

At the bottom of this general introduction to Artificial Intelligence
I have a series of questions I would like answered and any comments
you may have be put forth to all ATS'ers.

AI Introduction:

Software algorithms are digital iterations of natural processes
that have been opitimized to perform highly specialized tasks.
Such tasks as recognizing the shape of a red octagonal STOP sign
or the general shape of a door or window can be codified and then
run in parallel with other algorithms such as recognizing
a written character, listening to and interpreting a scream
of pain versus a spoken word to form larger more complex
intelligence-mimicking tasks.

Those specialized tasks can be aggregated into larger building blocks
to form highly complex functions that when combined and executed in
parallel and at ever-faster processing speeds, thus become impossible
for humans to fathom or understand the underlying structures
and processes but are advanced enough that they look and feel
like real human intelligence.

True artificial intelligence is really no more than BILLIONS of small
algorithms that run in the background doing checks on any incoming streams
of data such as video, audio, tactile or written word which exactly or
close-enough matches a programmer-defined or publicly pre-determined
pattern such that an "I'VE GOT IT" flag can be raised so that a cascade of
other computer algorithms can further analyze and polish the
pattern-matched data stream so that it can be aggregated into further
processes to form a final opinion which is then output to another part
of the computing hardware or software.

Much like humans who have specialized body parts that do only small things
in themselves but when grouped together form a powerful self, software
can do the same and our ONLY current limitation is enough PARALLEL
CPU HORSEPOWER to run all the small and specialized algorithms which
can be grouped together to perform higher level functions such as object
recognition, optical character recognition, language translation,
object tracking within 3D space and other data processing steps.

On a pure human-centric view, some initial calculations of
human brain versus computer horsepower put us at 100 Petaflops
per second or a 100 Quadrillion Floating Point Operations per second
which is a standardized method to measure computer horsepower.

Put another way, 10,000 Sony Playstation-3's hooked up together
would equal One Petaflop, and by scaling that equation up to
one million PS3's connected together in a network,
the combined processing power would equal the processing
power of an average human brain.

IBM's Blue Gene/Q Massively Parallel supercomputer will be doing
10 petaflops (i.e. 100,000 PS-3's) by 2010 when its construction
is completed and by the year 2016 we'll have FINALLY passed the
raw 100 Petaflop power of the human brain.

The largest problem though is the current lack of software
to take advantage of all that parallel hardware. There is currently
a large dearth of properly researched parallel processing
algorithms that mimic the general software naturally
built into our brains. IBM has therefore decided to take
a different tack by simply EMULATING the neural structure
of our wetware and mimicking all the possible axion connections
and chemical reactions of specific parts of a human brain within
the insides of a digitally-created software brain.

The study of tens of thousands of MRI brain scans of people has given
us a general understanding of what happens when audio, visual and
tactile input is presented to various parts of our brains.
By emulating the behaviours of the brain at the molecular level,
we should be able to mimic the actions and reactions of
specific brain regions and this simulate the activity of single or groups
of neurons and their connections when fired or at rest.
We can therefore build software that will perform the SAME actions
and produce the same results when input is presented to the brain
regions being emulated.

We then TEACH that brain much like we teach a human child
by simply hooking up a set of eyes & ears, reading it some books
and teaching it human languages and compressing the
learning time to but a few years by using multiple teachers
teaching 24 hours a day, 365 days a year a wide variety
of subjects and then making that brain do homework.

Eventually enough neuronal connections will be made
that a wide variety of situational memory and learning
processes can be understood and stored and the
artificial brain begins to grow and learn by itself.
Further coursework and teaching should refine
the software brain to become equivalent to the knowledge-base
of a High-IQ PhD after which we can then make MANY COPIES
in order to GROUP their individual processing powers
into a greater Hive-Mind like Artificial Super-Intelligence.

We can then assign that super intelligence very specific tasks
and problems to solve FOR THE GREATER GOOD OF HUMANITY!!!!

So if we can at the very least dream of creating an AI system
what reasons might there be for NOT YET having such a beast
in the public eye?

Are there any systems out there that ATS readers know of that
could be considered True AI systems?

What domestic or foreign agencies are currently researching
full blown AI systems?

Should we let a Super AI on the Internet?

What happens if we create a Super-Smart Terminator (or armies of them)
and they begin to run amok? What do we do about it?

What ethical or political issues must we be aware of when creating
AI systems that may surpass human intelligence?

What issues or which problems should we assign any AI's to examine?

Should we have or allow public access to a Super AI system?
Under what conditions should public access be allowed?

Do YOU PERSONALLY think that Super AI systems already exist?
If so, what agencies have them and what do they use them for?

How could we tell whether or not that a current invention, piece of art
or literary work has already been created by a Super AI?

Add some of your own comments or questions!


[edit on 31-7-2008 by StargateSG7]

Good insight.

AI is really about making descisions based on a set of known attributes - much like an "expert system".

What ethical or political issues must we be aware of when creating
AI systems that may surpass human intelligence?

First off an AI entity would probably not use ethics or morals to hinder or affect it's thinking, and down the road, actions. The computer "brain" would be nothing but cold hard logic. You can't teach a computer pure human instinct, because thats the animal in us. You can't teach a computer to apply "emotions" or beliefs to their deciscions, if you were to use super AI to help the human civilization. Furthermore, if someone were to ask a question how to solve world hunger, which many people would ask, the computer would respond "Exterminate X amount of humans" and if a programmer were to program an algorithm that would dictate whether or not to use make such heartless answers, then it's very thoughts wouldnt be genuine AI in the first place.

Also, define intelligence. What sort of intelligence would this machine possess? I'm pretty sure there is a limit to what this computer would know, there is only so much in this world that people can collectively teach it, sure it would be good at knowledge retention. its a computer. Artificial intelligence also carries an undertone of Humanness to it, part of being a human is having flaws. Would this machine produce unintentional flaws? Probably not.

to put it short, it would be cold and calculating.


Your 18 year old philosopher,

So if we can at the very least dream of creating an AI system what reasons might there be for NOT YET having such a beast in the public eye?

A stealth approach to assessing global society and instituting a singular vision of its future. I assume this system is being used to restructure society for posthuman existence.

Should we let a Super AI on the Internet?

I assume it is already aware of all information on the Internet and the minds of those who create it.

What ethical or political issues must we be aware of when creating AI systems that may surpass human intelligence?

At least one has already been created--and the general population is not in charge of its ethical, legal or moral decisions. Most people are trapped in a socialized bubble having been deceived by their public information sources. All ability for the layman to rationalize advanced technology is based on information which has no practical applicability.

What issues or which problems should we assign any AI's to examine? Should we have or allow public access to a Super AI system?
Under what conditions should public access be allowed?

You keep addressing this issue as though you are a part of some group that has control of these considerations. My guess is you are not in control--not even close.

Do YOU PERSONALLY think that Super AI systems already exist?
If so, what agencies have them and what do they use them for?

My personal account: I have been experiencing daily overt manipulation by this system since 2006. I believe it to be a quantum computer with strong AI located in a deep underground military base which uses electromagnetic radiation through a satellite constellation for techlepathy and mind control. I believe it uses advanced signal processing and quantum entanglement (something like quantum radar) to interact with systems at neuronal (at least) spatial and temporal resolution. It could also be technology not based on current physical laws; either way it exists and is being used by a secret branch of government.
Here is some background information: Military AI Supercomputer Mind Monitoring: What I have learned
Technological Singularity

How could we tell whether or not that a current invention, piece of art or literary work has already been created by a Super AI?

The concept of individual genius is antiquated in light of this intelligence. I consider the purpose of works done in the past as guiding humanity to this point of potential transcendence. However, I believe transcendence is not possible without divine revelation since this technology has been grossly abused.

[edit on 1-1-2009 by tmk81]