Computing machinery and intelligence

Abstract

The question of whether machines can think is functionally replaced by the “imitation game,” a behavioral test determining if a digital computer can mimic human responses closely enough to deceive an interrogator. Digital computers, characterized by their storage, executive units, and control systems, are universal discrete-state machines capable of mimicking any other discrete-state system. While various theological, mathematical, and philosophical objections challenge the possibility of machine intelligence, many of these rely on unfounded assumptions regarding human superiority or the nature of consciousness. Specifically, the mathematical limitations identified by Gödel and others do not necessarily prove human intellect is exempt from similar constraints. Achieving machine intelligence likely requires moving beyond rigid, exhaustive programming toward the development of learning machines. By simulating the initial state of a child’s mind and subjecting it to an appropriate course of education, a machine may acquire complex intellectual capacities through experience and reinforcement. This evolutionary approach to programming suggests that machines will eventually compete with humans in all purely intellectual fields. – AI-generated abstract.