An unethical optimization principle.

AI ethics artificial intelligence economics extreme value theory financial regulation

Journal

Royal Society open science
ISSN: 2054-5703
Titre abrégé: R Soc Open Sci
Pays: England
ID NLM: 101647528

Informations de publication

Date de publication:
Jul 2020
Historique:
received: 19 03 2020
accepted: 05 06 2020
entrez: 3 9 2020
pubmed: 3 9 2020
medline: 3 9 2020
Statut: epublish

Résumé

If an artificial intelligence aims to maximize risk-adjusted return, then under mild conditions it is disproportionately likely to pick an unethical strategy unless the objective function allows sufficiently for this risk. Even if the proportion

Identifiants

pubmed: 32874640
doi: 10.1098/rsos.200462
pii: rsos200462
pmc: PMC7428226
doi:

Types de publication

Journal Article

Langues

eng

Pagination

200462

Informations de copyright

© 2020 The Authors.

Déclaration de conflit d'intérêts

At the time of writing, R.S.M. is an associate editor of Royal Society Open Science, but he had no involvement in the review or assessment of the paper.

Références

Science. 2019 Nov 22;366(6468):999-1004
pubmed: 31754000

Auteurs

Nicholas Beale (N)

Sciteb Ltd, 23 Berkeley Square, London W1J 6HE, UK.

Heather Battey (H)

Department of Mathematics, Imperial College London, 180 Queen's Gate, London SW7 2AZ, UK.

Anthony C Davison (AC)

Institute of Mathematics, Ecole Polytechnique Fédérale de Lausanne, Station 8, 1015 Lausanne, Switzerland.

Robert S MacKay (RS)

Mathematics Institute, University of Warwick, Coventry CV4 7AL, UK.

Classifications MeSH