A causal perspective on AI deception in games
File(s)paper2CAUSAL.pdf (1.4 MB)
Published version
Author(s)
Ward, Francis
Toni, francesca
Belardinelli, francesco
Type
Conference Paper
Abstract
Deception is a core challenge for AI safety and we focus on the problem that AI agents might learn
deceptive strategies in pursuit of their objectives. We define the incentives one agent has to signal to
and deceive another agent. We present several examples of deceptive artificial agents and show that our
definition has desirable properties.
deceptive strategies in pursuit of their objectives. We define the incentives one agent has to signal to
and deceive another agent. We present several examples of deceptive artificial agents and show that our
definition has desirable properties.
Date Issued
2022-07-31
Date Acceptance
2022-06-03
Citation
CEUR-ART, 2022, pp.1-16
Publisher
CEUR Workshop Proceedings
Start Page
1
End Page
16
Journal / Book Title
CEUR-ART
Copyright Statement
© 2022 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
License URL
Identifier
https://ceur-ws.org/Vol-3193/paper2CAUSAL.pdf
Source
AI Safety 2022 (IJCAI-ECAI-22)
Publication Status
Published
Start Date
2022-07-24
Coverage Spatial
Vienna
Date Publish Online
2022-07-31