Constructivism and its risks in artificial intelligence

Abstract

The research and development of artificial intelligence (AI) technologies involve choices that extend well beyond the search for narrow engineering solutions to problems. The label ‘constructivism’ is used to capture this larger realm of social choice. Drawing on the history of AI, a distinction is made between limited artificial narrow intelligence (ANI) and artificial general intelligence (AGI). Both forms, the paper argues, carry risks. Following this history, the paper outlines how different approaches to rationality have led to different ‘tribes’ of AI. No universal model of rationality is available to AI engineers. Choice is everywhere. The paper then moves to an exploration of the links between AI and chess. It argues that chess, far from being an objective measure of rationality and intelligence, reveals the subjective biases and risks involved in the pursuit of AI. The paper moves on to provides examples of various unstable and potentially dangerous race heats taking place in AI, including those among various AI research groups (public and private), among corporations and among states. The final section draws together the various risks of AI.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 12 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.