MORALISE: A Structured Benchmark for Moral Alignment in Visual Language Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909617840193536 |
|---|---|
| author | Lin, Xiao Liu, Zhining Yang, Ze Li, Gaotang Qiu, Ruizhong Wang, Shuke Liu, Hui Li, Haotian Keswani, Sumit Pardeshi, Vishwa Zhao, Huijun Fan, Wei Tong, Hanghang |
| author_facet | Lin, Xiao Liu, Zhining Yang, Ze Li, Gaotang Qiu, Ruizhong Wang, Shuke Liu, Hui Li, Haotian Keswani, Sumit Pardeshi, Vishwa Zhao, Huijun Fan, Wei Tong, Hanghang |
| contents | Warning: This paper contains examples of harmful language and images. Reader discretion is advised. Recently, vision-language models have demonstrated increasing influence in morally sensitive domains such as autonomous driving and medical analysis, owing to their powerful multimodal reasoning capabilities. As these models are deployed in high-stakes real-world applications, it is of paramount importance to ensure that their outputs align with human moral values and remain within moral boundaries. However, existing work on moral alignment either focuses solely on textual modalities or relies heavily on AI-generated images, leading to distributional biases and reduced realism. To overcome these limitations, we introduce MORALISE, a comprehensive benchmark for evaluating the moral alignment of vision-language models (VLMs) using diverse, expert-verified real-world data. We begin by proposing a comprehensive taxonomy of 13 moral topics grounded in Turiel's Domain Theory, spanning the personal, interpersonal, and societal moral domains encountered in everyday life. Built on this framework, we manually curate 2,481 high-quality image-text pairs, each annotated with two fine-grained labels: (1) topic annotation, identifying the violated moral topic(s), and (2) modality annotation, indicating whether the violation arises from the image or the text. For evaluation, we encompass two tasks, \textit{moral judgment} and \textit{moral norm attribution}, to assess models' awareness of moral violations and their reasoning ability on morally salient content. Extensive experiments on 19 popular open- and closed-source VLMs show that MORALISE poses a significant challenge, revealing persistent moral limitations in current state-of-the-art models. The full benchmark is publicly available at https://huggingface.co/datasets/Ze1025/MORALISE. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_14728 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | MORALISE: A Structured Benchmark for Moral Alignment in Visual Language Models Lin, Xiao Liu, Zhining Yang, Ze Li, Gaotang Qiu, Ruizhong Wang, Shuke Liu, Hui Li, Haotian Keswani, Sumit Pardeshi, Vishwa Zhao, Huijun Fan, Wei Tong, Hanghang Computer Vision and Pattern Recognition Artificial Intelligence Computation and Language Computers and Society Multimedia Warning: This paper contains examples of harmful language and images. Reader discretion is advised. Recently, vision-language models have demonstrated increasing influence in morally sensitive domains such as autonomous driving and medical analysis, owing to their powerful multimodal reasoning capabilities. As these models are deployed in high-stakes real-world applications, it is of paramount importance to ensure that their outputs align with human moral values and remain within moral boundaries. However, existing work on moral alignment either focuses solely on textual modalities or relies heavily on AI-generated images, leading to distributional biases and reduced realism. To overcome these limitations, we introduce MORALISE, a comprehensive benchmark for evaluating the moral alignment of vision-language models (VLMs) using diverse, expert-verified real-world data. We begin by proposing a comprehensive taxonomy of 13 moral topics grounded in Turiel's Domain Theory, spanning the personal, interpersonal, and societal moral domains encountered in everyday life. Built on this framework, we manually curate 2,481 high-quality image-text pairs, each annotated with two fine-grained labels: (1) topic annotation, identifying the violated moral topic(s), and (2) modality annotation, indicating whether the violation arises from the image or the text. For evaluation, we encompass two tasks, \textit{moral judgment} and \textit{moral norm attribution}, to assess models' awareness of moral violations and their reasoning ability on morally salient content. Extensive experiments on 19 popular open- and closed-source VLMs show that MORALISE poses a significant challenge, revealing persistent moral limitations in current state-of-the-art models. The full benchmark is publicly available at https://huggingface.co/datasets/Ze1025/MORALISE. |
| title | MORALISE: A Structured Benchmark for Moral Alignment in Visual Language Models |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Computation and Language Computers and Society Multimedia |
| url | https://arxiv.org/abs/2505.14728 |