The notion of intelligent automatons and calculating devices goes back to antiquity.
The ancient Greeks circa 150BC invented the Antikythera mechanism – the first known analogue ship navigation calculator – More Info It was not until the 15th century that the metal smithing technology used to manufacture clocks approached the skill exhibited by the makers of the Antikythera.
In the 18th century chess playing automaton called The Turk were bogus frauds where the game was played by a person inside the machine.
In the Victorian era of the 19th century, work by Charles Babbage attempted to model human mathematical intelligence in the form of mechanical calculators which led to the invention of electronic computers.
In 1948, two giants in the early development of electronic computers Alan Turing and John von Neumann debated a core notion of what is an intelligent machine.
Essentially if a so called intelligent machine is missing a feature, the we, the human developers of the machine can always build the new functionality into a revised version of the intelligent machine. In other words, intelligent machines can only solve problems for which they are designed and built for.
We can construct sensors that see, hear, talk, feel or smell, mechanical devices that grip – all manner of computer technology that could be used in creating digital automatons but at the end of the day, the systems can only perform the tasks defined in the device’s use case architecture. No use case – no function point.
All living things have some form of intelligence built into them while higher organisms can reason and solve for new problems.
A video of a Pepino the octopus screwing the lid off a glass jar proves to be the right trick to catch a crab. It demonstrates what intelligence means where the octopus used a reasoning process to figure out that
a) the crab was encased in a transparent container and
b) the lid was removed to get the crab into the jar and
c) removing the lid is the key route to a tasty snack.
How to remove the lid was not a skill learned in the wild.
The winning solution was for the octopus is to:
- Hold the jar still with two tentacles
- Grasp the lid with four tentacles.
- A:try turning the jar clockwise with the jar grasping tentacles while tightly grasping the lid.
- B: If this fails, try and turning the jar counter clockwise.
- C: Pop off the lid after sensing that the lid is coming off.
Pepolino clearly demonstrates the ability to formulate a goal and solve a problem never encountered before.
The comments by von Neumann hits home that the computer can only be as intelligent as its design and implementation. We can add new features to the device but in the end its functionality is constrained by the scope of use cases implemented within the system. The term “Artificial Intelligence” or AI then a misnomer.
I postulate that AI just another implementation of computer technology.
It would never be able to figure out on its own how to catch the crab inside a jar unless the use case was implemented in an electronic octopus.
All IT systems are developed with three constraints: cost, time and labour.
All intelligent systems being constructed are subject to these three constraints.
Many projects will be experimental where the end solution cannot be specified where no prior experience can be relied upon to guide the development. An experimental prototype would be called for in this case. The final solution won’t come cheap.
This almost guarantees that whatever is created will be limited to the three constraints.
The octopus on the other hand has had 300 million years to evolve and adapt using a trial and error evolutionary approach to enable it to camouflage its body with the environment, gain excellent eyesight and the smarts to find food inside a glass jar.