Record STMT-W243F60Voluntary

Robots Can Feel: LLM-based Framework for Robot Ethical Reasoning

Framework Institutional policyEastern Europe · RUproposed

This paper presents 'Robots Can Feel', a novel ethical reasoning framework for robots that combines logic with human-like emotion simulation to make decisions in morally complex situations. The system utilizes an adjustable Emotion Weight Coefficient to manage the role of emotions in robot decision-making and was tested across eight top-tier large language models. The evaluation demonstrated that the Emotion Weight Coefficient significantly influences the robot's final decision in morally challenging scenarios such as animal compassion and dietary violations.

Summary and content fields are drawn from the document text and are pending editorial review.

Who issued it

Skolkovo Institute of Science and TechnologyAcademic · RU

Lead issuer

arXivAcademic

Publisher

Legal forceVoluntary

Under whose authorityInstitutional policy

Judged fromfull text

Who it governs

Researchers and academicsCompaniesAI developers and providersAI deployers and users

AI researchers and roboticists interested in endowing cognitive robots with human-like ethical decision-making capabilities.

Sectors

Research and science · cognitive robotics

What the document contains

What's covered

The document covers the architecture and empirical evaluation of an LLM-based ethical reasoning module for cognitive robots that simulates human-like emotions alongside logical rules.

What's missing

The text does not provide formal legal enforcement mechanisms, industry safety certifications, or real-world deployment safety guarantees.

Key takeaways

The proposed 'Robots Can Feel' framework combines logical analysis and simulated human emotions for robot ethical reasoning.
An Emotion Weight Coefficient (EWC) allows users to calibrate the influence of emotions on robot decision-making from 0 to 1.
Evaluation across eight prominent large language models showed that the EWC consistently alters decision outcomes across different base models.
ANOVA analysis confirmed statistically significant differences in final decisions when applying varying EWCs in animal compassion and dietary scenarios.

Outline

· I. Introduction
· II. Related Works
· III. Robotic Ethical Reasoning
· IV. System Overview
· V. Evaluation
· VI. Conclusion

Issued9 May 2024

Length4,638 words · 6 pages

FormatNot recorded

Scopenational

How it scores

Pending review

Each score is backed by a quote verified against the document text. Editorial review has not yet confirmed them.

Accountability

0/5

Clear responsibility for AI outcomes

screening score · 2026-10-24

Fairness

0/5

Non-discrimination and equitable treatment

screening score · 2026-10-24

Human Dignity

0/5

Respect for the inherent worth of persons

screening score · 2026-10-24

Human Oversight

0/5

Humans remain in control of consequential decisions

screening score · 2026-10-24

Privacy

0/5

Protection of personal data and private life

screening score · 2026-10-24

Transparency

1/5

Openness about how AI systems work and are used

For ease of formatting and further parsing of the robot’s ethical reasoning, it is represented in XML format.

IV. SYSTEM OVERVIEW

screening score · 2026-10-24

Enforcement

0/5

Consequences and compliance mechanisms with real force

screening score · 2026-10-24

Impact Assessment

3/5

Required assessment of an AI system's risks/impacts before or during use

Within the proposed CognitiveOS, an ethical assessment mechanism is employed, leveraging Retrieval- Augmented Generation (RAG) technology.

II. RELATED WORKS

screening score · 2026-10-24

Topics

Physical AI Governancesupporting

robotics ethics large_language_models artificial_intelligence cognitive_robotics