Header

UZH-Logo

Maintenance Infos

Human control redressed: comparing AI and human predictability in a real-effort task


Kandul, Serhiy; Micheli, Vincent; Beck, Juliane; Burri, Thomas; Fleuret, François; Kneer, Markus; Christen, Markus (2023). Human control redressed: comparing AI and human predictability in a real-effort task. Computers in Human Behavior Reports, 10:100290.

Abstract

Predictability is a prerequisite for effective human control of artificial intelligence (AI). The inability to predict malfunctioning of AI, for example, impedes timely human intervention. In this paper, we empirically investigate how AI’s predictability compares to the predictability of humans in a real-effort task. We show that humans are worse at predicting AI performance than at predicting human performance. Importantly, participants are not aware of the differences in relative predictability of AI and overestimate their prediction skills. These results raise doubts about the human ability to effectively exercise control of AI — at least in certain contexts.

Abstract

Predictability is a prerequisite for effective human control of artificial intelligence (AI). The inability to predict malfunctioning of AI, for example, impedes timely human intervention. In this paper, we empirically investigate how AI’s predictability compares to the predictability of humans in a real-effort task. We show that humans are worse at predicting AI performance than at predicting human performance. Importantly, participants are not aware of the differences in relative predictability of AI and overestimate their prediction skills. These results raise doubts about the human ability to effectively exercise control of AI — at least in certain contexts.

Statistics

Citations

Dimensions.ai Metrics

Altmetrics

Downloads

19 downloads since deposited on 30 Jun 2023
20 downloads since 12 months
Detailed statistics

Additional indexing

Item Type:Journal Article, refereed, original work
Communities & Collections:01 Faculty of Theology and the Study of Religion > Center for Ethics
06 Faculty of Arts > Institute of Philosophy
04 Faculty of Medicine > Institute of Biomedical Ethics and History of Medicine
Dewey Decimal Classification:100 Philosophy
Scopus Subject Areas:Life Sciences > Neuroscience (miscellaneous)
Social Sciences & Humanities > Applied Psychology
Physical Sciences > Human-Computer Interaction
Physical Sciences > Computer Science Applications
Life Sciences > Cognitive Neuroscience
Physical Sciences > Artificial Intelligence
Uncontrolled Keywords:Artificial Intelligence, Cognitive Neuroscience, Computer Science Applications, Human-Computer Interaction, Applied Psychology, Neuroscience (miscellaneous), Human Control, AI Predictability, Lunar Lander Game, Human-Computer Interaction
Language:English
Date:3 May 2023
Deposited On:30 Jun 2023 15:03
Last Modified:09 Feb 2024 08:33
Publisher:Elsevier
ISSN:2451-9588
OA Status:Gold
Publisher DOI:https://doi.org/10.1016/j.chbr.2023.100290
Project Information:
  • : FunderSNSF
  • : Grant ID187494
  • : Project TitleMeaningful Human Control of Security Systems - Aligning International Humanitarian Law with Human Psychology
  • Content: Published Version
  • Language: English
  • Licence: Creative Commons: Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0)