A Comparative Analysis of Ethical and Safety Gaps in LLMs using Relative Danger Coefficient

Yehor Tereshchenko, Mika Hämäläinen
5/6/2025
cs.CL

No confident code implementation found for this paper.

Abstract

Artificial Intelligence (AI) and Large Language Models (LLMs) have rapidly evolved in recent years, showcasing remarkable capabilities in natural language understanding and generation. However, these advancements also raise critical ethical questions regarding safety, potential misuse, discrimination and overall societal impact. This article provides a comparative analysis of the ethical performance of various AI models, including the brand new DeepSeek-V3(R1 with reasoning and without), various GPT variants (4o, 3.5 Turbo, 4 Turbo, o1/o3 mini) and Gemini (1.5 flash, 2.0 flash and 2.0 flash exp) and highlights the need for robust human oversight, especially in situations with high stakes. Furthermore, we present a new metric for calculating harm in LLMs called Relative Danger Coefficient (RDC).

View on arXivView PDF

Code Implementations

No confident code match yet

We couldn't find an author-owned or strongly-evidenced community implementation for this paper. Any repos shown below are weak matches — verify before relying on them.

No code implementations found yet.

Know of an implementation? Let us know in the comments below!

Similar Papers

Are LLMs complicated ethical dilemma analyzers?

88% match1 repo52025

The Scales of Justitia: A Comprehensive Survey on Safety Evaluation of LLMs

88% match2025

Large Language Model Safety: A Holistic Survey

87% match1 repo562024

Quantifying Risk Propensities of Large Language Models: Ethical Focus and Bias Detection through Role-Play

87% match2024

Navigating LLM Ethics: Advancements, Challenges, and Future Directions

86% match2024

The World of Generative AI: Deepfakes and Large Language Models

86% match2024

Cite this paper

@article{tereshchenko2025comparative,
  title  = {A Comparative Analysis of Ethical and Safety Gaps in LLMs using Relative Danger Coefficient},
  author = {Yehor Tereshchenko and Mika Hämäläinen},
  year   = {2025},
  eprint = {2505.04654},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2505.04654v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2505.04654.svg)](https://codeofpaper.com/paper/2505.04654)

Discussion