Empirical Foundations for Meaningful Human Control:
Keywords
Loading...
Authors
Issue Date
2024-08-01
Language
en
Document type
Journal Title
Journal ISSN
Volume Title
Publisher
Title
ISSN
Volume
Issue
Startpage
Endpage
DOI
Abstract
The rise of autonomous artificial intelligence (AI) systems has put strain on traditional notions of
control. The inherent complexity of AI processes leads to an opaqueness which presents challenges
in monitoring actions and attributing responsibility. These challenges may be referred to as ‘control
gaps’ or ‘responsibility gaps’. To address these, recent theories of Meaningful Human Control (MHC)
have emerged, with the aim of establishing clear requirements for effective human oversight of AI
systems. In doing so, some accounts of MHC relax the need for direct operational relationships
between human operators and the systems under their control.
This paper advances a novel philosophical thesis: We assert that the Sense of Agency—a cognitive
measure closely associated with control—can empirically substantiate these normative foundations
of MHC. Through rigorous analysis of scientific literature on the Sense of Agency, we construct a
compelling argument for the empirical justification of the philosophical theory of action underlying
MHC.
The notion that reasons and intentions play an exceedingly important role in human control,
previously considered primarily within theoretical discourse, we find has significant support in recent
empirical studies of human cognition. This connection not only provides crucial empirical grounding
to existing MHC theories, but it also opens avenues for design insights from both fields.
In our examination of MHC and the Sense of Agency, we identify a crucial intersection between these
concepts, so far underexplored, which we propose holds significant potential for future inquiry and
practical development.
Description
Citation
Supervisor
Faculty
Faculteit der Sociale Wetenschappen