AI model collapse might be prevented by studying human language transmission

Kenny Smith; Simon Kirby; Shangmin Guo; Thomas L. Griffiths

doi:10.1038/d41586-024-03023-y

AI model collapse might be prevented by studying human language transmission

Kenny Smith
, Simon Kirby
, Shangmin Guo
, Thomas L. Griffiths

Research output: Contribution to journal › Comment/debate › peer-review

4 Scopus citations

Abstract

Letter

Original language	English (US)
Pages (from-to)	525
Number of pages	1
Journal	Nature
Volume	633
Issue number	8030
DOIs	https://doi.org/10.1038/d41586-024-03023-y
State	Published - Sep 19 2024

All Science Journal Classification (ASJC) codes

General

Keywords

Culture
Language
Machine learning

Access to Document

10.1038/d41586-024-03023-y

Cite this

@article{4dd99828ca35431abf2c5d37bef6479c,

title = "AI model collapse might be prevented by studying human language transmission",

abstract = "Letter",

keywords = "Culture, Language, Machine learning",

author = "Kenny Smith and Simon Kirby and Shangmin Guo and Griffiths, \{Thomas L.\}",

note = "Publisher Copyright: {\textcopyright} Springer Nature Limited 2024.",

year = "2024",

month = sep,

day = "19",

doi = "10.1038/d41586-024-03023-y",

language = "English (US)",

volume = "633",

pages = "525",

journal = "Nature",

issn = "0028-0836",

publisher = "Nature Research",

number = "8030",

}

TY - JOUR

T1 - AI model collapse might be prevented by studying human language transmission

AU - Smith, Kenny

AU - Kirby, Simon

AU - Guo, Shangmin

AU - Griffiths, Thomas L.

N1 - Publisher Copyright: © Springer Nature Limited 2024.

PY - 2024/9/19

Y1 - 2024/9/19

N2 - Letter

AB - Letter

KW - Culture

KW - Language

KW - Machine learning

UR - https://www.scopus.com/pages/publications/85204380731

UR - https://www.scopus.com/pages/publications/85204380731#tab=citedBy

U2 - 10.1038/d41586-024-03023-y

DO - 10.1038/d41586-024-03023-y

M3 - Comment/debate

C2 - 39289495

AN - SCOPUS:85204380731

SN - 0028-0836

VL - 633

SP - 525

JO - Nature

JF - Nature

IS - 8030

ER -

AI model collapse might be prevented by studying human language transmission

Abstract

All Science Journal Classification (ASJC) codes

Keywords

Access to Document

Other files and links

Fingerprint

Cite this