Visual Distraction Undermines Moral Reasoning in Vision-Language Models
This exposes critical fragilities in multimodal AI safety, highlighting an urgent need for alignment beyond text-only contexts, which is incremental as it builds on existing moral evaluation benchmarks.
The study tackled the problem of ensuring consistent moral reasoning in AI systems as they evolve to handle visual inputs, finding that visual distractions in Vision-Language Models bypass text-based safety mechanisms and alter moral decision-making, with evaluations showing vision activates intuition-like pathways that override safer reasoning patterns.
Moral reasoning is fundamental to safe Artificial Intelligence (AI), yet ensuring its consistency across modalities becomes critical as AI systems evolve from text-based assistants to embodied agents. Current safety techniques demonstrate success in textual contexts, but concerns remain about generalization to visual inputs. Existing moral evaluation benchmarks rely on textonly formats and lack systematic control over variables that influence moral decision-making. Here we show that visual inputs fundamentally alter moral decision-making in state-of-the-art (SOTA) Vision-Language Models (VLMs), bypassing text-based safety mechanisms. We introduce Moral Dilemma Simulation (MDS), a multimodal benchmark grounded in Moral Foundation Theory (MFT) that enables mechanistic analysis through orthogonal manipulation of visual and contextual variables. The evaluation reveals that the vision modality activates intuition-like pathways that override the more deliberate and safer reasoning patterns observed in text-only contexts. These findings expose critical fragilities where language-tuned safety filters fail to constrain visual processing, demonstrating the urgent need for multimodal safety alignment.