Artificial Intelligence has been advancing at an ever quickening pace. It is mostly going unnoticed by the average person. I guess it depends how one defines artificial intelligence. What one person may call AI(Artificial Intelligence), another may not. For example I think that on some basic level the newest digital cameras contain AI. After all they decide what you aperture should be, how fast the shutter speed is set to, distance to the object and many other things. It seems to me that this takes some sort of rudimentary electronic intelligence. True it might be less than that of an ameba but the camera is making decisions. We could look at chess machines and say they had intelligence but many might say that they simply use mathematics and preprogrammed moves as a basis for their move. But aren't we preprogrammed to know what to do in certain situations? That is why we go to school, and why our parents teach us the basics of life when we are children. If we were raised by wolves, and discovered when we were 20 years old and brought back to civilization, would we know how to interact with other people? Of course not.

The point I am trying to make is that if we are trained to respond a certain way (programmed) and a computer is programmed to respond a certain way, why can't an advanced machine be called intelligent?

Some think that it is not enough that a machine can make an intelligent decision, they feel that the decision must be arrived at in the same way a human arrives at a decision. This has led to the development of neural networks for computers. The decision making in these machines is modeled after the decision making process in the human brain. But other people think the only limiting factor to making machines intelligent is the speed of the processor. They claim that 100 billion MIPS (1 MIPS = 1 Million Instructions Per Second) will be needed before a machine can react as fast as a human brain. We are not talking of just calculations here, as we know machines are much faster than we are at them. We are talking of machines having to consider the myriad of things involved with an intelligent decision. We are getting near that speed point now. In the year 2005 a machine from IBM is scheduled to go online that will be capable of 1 billion MIPS. We know that computing power doubles in 18 months to two years. This means that if we use 24 months as the base for doubling the speed of processors, in 2019 we will have reached 128 billion MIPS. This is not taking into account switching to computers that run on light which would give us an approximate speed gain of 8000 times or Quantum computers which would also increase processing speed.

The second part of the problem of creating a machine that will work as a human brain is building the correct software. We don't want a machine to turn out to be a monster because some software manufacturer made several errors in the programming. Can you imagine a patch coming out that would do away with the homicidal tendencies of you personal robot? Aside from this, how do you write a program to mimic a brain that nobody understands? Just for fun lets say that by 2019 we completely understand the working of the human brain. We write the program and the hardware has the speed. It works, but do we have a living being? Here are some of the things that determine if something is alive:

It is organized into complex structures based on organic molecules.
It maintains some type of homeostasis.
It grows and develops.
It reproduces and passes on genetic material as a blueprint for growth and subsequent reproduction
It acquires matter and energy from the external environment and converts it into different forms.
It responds to stimuli from the environment.
It evolves.

The time may come when the definition of life might have to change. Imagine if we came across a race of machines on a planet we were exploring that had the ability to communicate intelligently with us, would we consider them life?

Artificial Intelligence has been with us for some time now as you will see when you look at Spacelab Deutsche 2 below. The picture was taken in 1993. In the sixties and seventies it had been studied and attempts were made at developing it but they were severely hampered by the available hardware, not to mention the lack of software. But that was not the beginning of the search for artificial intelligence. In China it was reported that some sort of mechanical man was build of jade in the 3 or 4 century B.C. along with all sorts of mechanical birds that sang, fish that swam and dolls of men that moved. All through out history one finds all sorts of automata. It seems the search or AI is actually thousands of years old.

NASA is starting to build intelligence into its satellites and computer programs.

NASA Robot - NASA is hoping that this robot can someday go on missions too dangerous for humans. Source: NASA

Below is a short description of ASPEN a NASA project.

The following is from NASA JPL:

The Artificial Intelligence Group at JPL has been working on a system called ASPEN (A Scheduling and Planning Environment). Based on AI techniques, ASPEN is a modular, reconfigurable application framework which is capable of supporting a wide variety of planning and scheduling applications. ASPEN provides a set of reusable software components that implement the elements commonly found in complex planning/scheduling systems, including: an expressive modeling language, a resource management system, a temporal reasoning system, a set of search algorithms for generating and repairing schedules, and a graphical interface.

The primary application area for ASPEN is the spacecraft operations domain. Planning and scheduling spacecraft operations involves generating a sequence of low-level spacecraft commands from a set of high-level science and engineering goals. ASPEN will encode complex spacecraft operability constraints, flight rules, spacecraft hardware models, science experiment goals and operations procedures to allow for automated generation of low level spacecraft sequences by use of advanced Artificial Intelligence planning and scheduling technology.

New Millennium EO-1 is a earth imaging satellite featuring an advanced multi-spectral imaging device. EO-1 mission operations consists of managing spacecraft operability constraints (power, thermal, pointing, buffers, consumables, telecommunications, etc.) and science goals (imaging of specific targets within particular observation parameters). Of particular difficulty is managing the downlinks as the amount of data generated by the imaging device is quite large and ground contacts are a limited resource. In addition, because science targets for EO-1 are based upon short-term cloud predictions, schedules must be generated daily. Without an automated planning system such as ASPEN, daily scheduling would not be possible with EO-1’s three person mission operations team.

Automated planning/scheduling technologies have great promise in reducing operations cost and increasing the autonomy of the EO-1 mission operations. By automating the command sequence generation process and by encapsulating the operation specific knowledge, ASPEN will enable EO-1 spacecraft commanding by a small operations team without subsystem experts.