Since the 1950s, the concept that a Machine may soon match the full extent and level of achievement of human intelligence has been met with hype and frenzy. We’ve now developed devices that can handle specific, narrow issues — “clever” Machine that can diagnose sickness, drive vehicles, interpret speech, even beat people at chess — but general intelligence remains elusive. So let’s get one thing straight: advances in Machine intelligence will not result in unstoppable computer-led revolutions.
They may modify the types of jobs individuals do, but they will not put humanity on the verge of extinction. There will be no apocalypse of robots. Intelligence testing and computational intelligence techniques have focused on well-structured and formal tasks. That is problems with a specific purpose and a limited number of solutions. Humans, on the other hand, are inventive, illogical, and inconsistent. Concentrating on these well-structured issues could be like seeking your lost keys in the brightest light.
There are additional issues that are more typical of human intelligence and should be investigated further. So-called insight problems are one type of problem. A step-by-step approach, such as an algorithm, usually cannot address insight problems, or if it does, the process is exceedingly time-consuming. On the other hand, Insight problems are marked by a reorganization of the solver’s approach to the problem. For example, a representation is given to the solver in path problems, including a beginning state, a goal state, and a collection of tools or operations that can move through the representation. None of them are offered to the solver in insight problems.