Quasi-Moral Agency and Structural Attribution in Recurrent Neural Networks: Foundations for Distributed Ethics

This article defends the thesis that certain types of software systems—particularly those used in procedures with real-life moral implications, such as those implementing recurrent neural networks (RNNs, LSTMs, GRUs)—should be understood as structural participants in human moral activity. As such, they qualify as quasi-moral agents appropriate for functional normative assessment. This status depends not on consciousness or intentionality, but on their structural capacity to produce auditable, intelligible, and causally relevant outcomes. Starting from a reinterpretation of attribution theory from a sociopsychological perspective, this article examines how technical mechanisms such as integrated gradients, attention layers, and layered relevance propagation can support forms of practical justification. From a metaethical perspective, it is argued that structural attribution allows these systems to be integrated into human normative practices, particularly in morally sensitive contexts such as the allocation of resources for social benefits, early detection of suicide risk on digital platforms, or content moderation in democratic processes. The article addresses the conceptual and technical limitations of this approach including the risks of functional anthropomorphism or the interpretive weaknesses of current attribution methods and the need to preserve human responsibility. Finally, the article concludes that the ethical governance of AI systems must include performance metrics as well as a structured design for normative accountability.

Authors

Publication Details

Journal
Open Insight
Published
2026-11-14
DOI
https://doi.org/10.23924/oi.v17i41.757
Primary Topic
Ethics and Social Impacts of AI
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Quasi-Moral Agency and Structural Attribution in Recurrent Neural Networks: Foundations for Distributed Ethics

Fabio Morandín Ahuerma
Open Insight
Ethics and Social Impacts of AI
article

Quasi-Moral Agency and Structural Attribution in Recurrent Neural Networks: Foundations for Distributed Ethics

Fabio Morandín Ahuerma
article en

Abstract

This article defends the thesis that certain types of software systems—particularly those used in procedures with real-life moral implications, such as those implementing recurrent neural networks (RNNs, LSTMs, GRUs)—should be understood as structural participants in human moral activity. As such, they qualify as quasi-moral agents appropriate for functional normative assessment. This status depends not on consciousness or intentionality, but on their structural capacity to produce auditable, intelligible, and causally relevant outcomes. Starting from a reinterpretation of attribution theory from a sociopsychological perspective, this article examines how technical mechanisms such as integrated gradients, attention layers, and layered relevance propagation can support forms of practical justification. From a metaethical perspective, it is argued that structural attribution allows these systems to be integrated into human normative practices, particularly in morally sensitive contexts such as the allocation of resources for social benefits, early detection of suicide risk on digital platforms, or content moderation in democratic processes. The article addresses the conceptual and technical limitations of this approach including the risks of functional anthropomorphism or the interpretive weaknesses of current attribution methods and the need to preserve human responsibility. Finally, the article concludes that the ethical governance of AI systems must include performance metrics as well as a structured design for normative accountability.

Open InsightVol. 17(41)
Ethics and Social Impacts of AI
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon
Quasi-Moral Agency and Structural Attribution in Recurrent Neural Networks: Foundations for Distributed Ethics — Fabio Morandín Ahuerma · Open Insight (2026) | TGRS Research Map | TGRS