The Psychology of Thinking in Creating AI

Abstract

The broad-scale emergence of AI in industry calls forth basic questions in terms of the knowledge bases and approaches relevant for its design. Engineering design has been mainly developed for electromechanical artifacts. In practice, this has meant that the scientific knowledge required for creating technical artifacts such as engines, cars, ships, cranes, telephones, radios, TVs, and simple data processing units has been natural science. However, one cannot find intelligent processes by means of physics and chemistry. Natural scientific phenomena follow their deterministic laws, but intelligence is based on selection and decision processes. The conceptual landscape of natural science is optimized for different kinds of phenomena than intelligent information processing. Consequently, the basic research under technology design should be rethought in light of the emergence of AI. To grasp intelligent information processing, we need concepts and approaches suited for the task. To create intelligent technologies on this basis, we need concepts and approaches that afford the operationalization of such information in artificial systems. We suggest that psychology of thinking and cognitive modeling provide a logical basis for future AI design.