Types of artificial intelligence

The scientific community distinguishes four types of artificial intelligence.

Type 1: machines whose algorithms know how to solve narrow tasks at a given moment in time. These machines are not capable of operating on data from the past and look for a solution to each problem anew.

The most prominent example of the first type of AI is IBM’s Deep Blue supercomputer, which managed to beat Garry Kasparov in chess in the late 1990s. Many of today’s machines are more complex, using neural networks, but they have not strayed far from the mechanical systems of the end of the last century.

Type 2: systems with limited memory.

An example of this system is the autopilot mechanism in a car, which can assess the situation based on recent experience. This system collects data about road markings, signs, and surroundings. It then analyses the information and makes decisions in real-time based on the traffic situation. The main disadvantage of such a system is the inability to store information forever, to develop permanent algorithms and habits, and to improve skills. The system works strictly according to algorithms and there is no way the driver can replace himself with an autopilot.

Type 3 and 4: machines that have self-awareness and initial intelligence. These systems will not only be able to learn from past experiences and predict solutions for future situations but also understand their purpose and mission.