Do VLMs Have a Moral Backbone? A Study on the Fragile Morality of Vision-Language Models (2026.findings-acl)
Copied to clipboard
Zhining Liu, Tianyi Wang, Xiao Lin, Penghao Ouyang, Gaotang Li, Ze Yang, Hui Liu, Sumit Keswani, Vishwa Pardeshi, Huijun Zhao, Wei Fan, Hanghang Tong
| Challenge: | Vision-Language Models (VLMs) have advanced multimodal learning, driving progress in cross-modal reasoning. |
| Approach: | They propose to examine moral robustness of vision-language models by analyzing their moral stances under multimodal perturbations. |
| Outcome: | The proposed model-agnostic multimodal perturbations expose VLMs to a variety of moral vulnerabilities, including a sycophancy trade-off where stronger instruction-following models are more susceptible to persuasion. |
Similar Papers
Tales of Morality: Comparing Human- and LLM-Generated Moral Stories from Visual Cues (2025.findings-emnlp)
Copied to clipboard
| Challenge: | a recent study has found that stories are central to how humans communicate moral values . |
| Approach: | They compare human- and LLM-generated moral narratives based on images annotated by humans for moral content . authors propose a framework for evaluating moral storytelling in vision-language models . |
| Outcome: | The proposed model compared human- and LLM-generated narratives on images . human stories reflect a balanced distribution of moral foundations and coherent narrative arcs, but LLMs emphasize Care foundation and lack emotional resolution. |
Beyond Words: Exploring Cultural Value Sensitivity in Multimodal Models (2025.findings-naacl)
Copied to clipboard
| Challenge: | Using large vision-language models to understand cultural contexts is a critical area of research. |
| Approach: | They conduct a thorough evaluation of multimodal models at different scales, focusing on their alignment with cultural values. |
| Outcome: | The proposed models show that they exhibit sensitivity to cultural values but their performance is highly context-dependent. |
Unveiling the Lack of LVLM Robustness to Fundamental Visual Variations: Why and Path Forward (2025.findings-acl)
Copied to clipboard
| Challenge: | Large Vision Language Models (LVLMs) have shown impressive performance on various vision-language tasks. |
| Approach: | They propose a benchmark framework for evaluating Visual Variation Robustness of Large Vision Language Models that incorporates automated evaluation dataset generation and principled metrics for thorough robustness assessment. |
| Outcome: | The proposed framework identifies a vulnerability to visual variations affecting even advanced models that excel at complex vision-language tasks but significantly underperform on simple tasks like object recognition. |
Visual Self-Fulfilling Alignment: Shaping Safety-Oriented Personas via Threat-Related Images (2026.acl-long)
Copied to clipboard
| Challenge: | Existing methods require explicit safety labels or contrastive data, yet visual inputs enable harmful outputs. |
| Approach: | They propose a visual self-fulfilling alignment mechanism that fine-tunes vision-language models on neutral VQA tasks without any safety labels. |
| Outcome: | The proposed approach reduces attack success rate, improves response quality, and mitigates over-refusal while preserving general capabilities. |
VIGNETTE: Socially Grounded Bias Evaluation for Vision-Language Models (2026.acl-long)
Copied to clipboard
| Challenge: | Existing studies on VLM bias focus on portrait-style images and gender-occupation associations . existing studies ignore broader and more complex social stereotypes and their implied harm . |
| Approach: | They propose a large-scale VQA benchmark for evaluating bias in vision-language models . they use a question-answering framework that spans factuality, perception, stereotyping, and decision making . |
| Outcome: | The proposed framework examines bias in vision-language models using 30M+ images . findings reveal subtle, multifaceted, and surprising stereotypical patterns . |
Do Emotions Influence Moral Judgment in Large Language Models? (2026.findings-acl)
Copied to clipboard
| Challenge: | Recent systems enforce explicit ethical constraints, but moral judgment rarely involves such clear-cut prohibitions. |
| Approach: | They develop an emotion-induction pipeline that infuses emotion into moral situations and evaluate shifts in moral acceptability across datasets and LLMs. |
| Outcome: | The proposed pipeline can infuses emotion into moral situations and evaluate moral acceptability shifts across datasets and LLMs. |
SaGE: Evaluating Moral Consistency in Large Language Models (2024.lrec-main)
Copied to clipboard
| Challenge: | Existing studies on Large Language Models (LLMs) have focused on accuracy but lack universally agreed-upon answers for moral scenarios. |
| Approach: | They propose a measure called Semantic Graph Entropy to measure a model's moral consistency grounded in "Rules of Thumb" they construct a moral Consistency Corpus (MCC) with 50K moral questions and the RoTs they followed to investigate LLM consistency on two popular datasets. |
| Outcome: | The proposed measure measures moral consistency on two popular datasets . |
Do Images Speak Louder than Words? Investigating the Effect of Textual Misinformation in VLMs (2026.eacl-long)
Copied to clipboard
| Challenge: | Existing studies have shown that Vision-Language Models have robust multimodal reasoning capabilities, but their robustness against textual misinformation remains under-explored. |
| Approach: | They propose to use visual-question-answering (VQA) prompts to generate persuasive prompts that deliberately conflict with visual evidence to test their models. |
| Outcome: | The proposed framework shows that models are vulnerable to misleading prompts, and show an average performance drop of over 48.2% after only one round of persuasive conversation. |
Structured Moral Reasoning in Language Models: A Value-Grounded Evaluation Framework (2025.emnlp-main)
Copied to clipboard
| Challenge: | Large language models (LLMs) are increasingly deployed in domains requiring moral understanding, yet their reasoning often remains shallow and misaligned with human reasoning. |
| Approach: | They propose a value-grounded framework for evaluating and distilling structured moral reasoning in large language models. |
| Outcome: | The proposed framework evaluates 12 open-source models across four moral datasets. |
Can VLMs Actually See and Read? A Survey on Modality Collapse in Vision-Language Models (2025.findings-acl)
Copied to clipboard
| Challenge: | Vision-language models integrate textual and visual information, enabling them to process visual inputs and generate predictions. |
| Approach: | They review work on modality collapse analysis to provide insights into the reason for this unintended behavior and review probing studies for fine-grained vision-language understanding. |
| Outcome: | The proposed models can achieve competitive performance in vision-language tasks despite relying heavily on textual information and ignoring visual information. |