AI is Dead: Long Live AI

Reference

Computer Science Technical Reports 166 (1999)

Degree Grantor

Abstract

The time seems ripe to reflect on the state of AI and re-evaluate its goals and techniques. It is argued that a fundamental oversight has been made in adopting a focus on knowledge and reasoning without simultaneously emphasising understanding and meaning. Despite not being able to define the latter term precisely, we can exploit our limited knowledge of meaning that most AI efforts will simply never achieve artificial intelligence of the kind that many pioneers in the field envisaged, or that most members of the public imagine when they hear this term. The only route to this goal is argued to be via learning, and objections to this approach are argued to be weak or flawed. A number of implications of learning with the specific aim of acquiring meaning and understanding capabilities are discussed.

Description

DOI

Related Link

Keywords

ANZSRC 2020 Field of Research Codes