Does Unification Come at a Cost? Uni-SafeBench: A Safety Benchmark for Unified Multimodal Large Models
This addresses safety risks in unified multimodal AI models, which is crucial for safer AGI development, though it is incremental as it builds on existing safety benchmarking.
The paper tackles the safety challenges of Unified Multimodal Large Models (UMLMs) by introducing Uni-SafeBench, a benchmark that evaluates safety across six categories and seven task types, revealing that unification degrades inherent safety and open-source UMLMs perform worse than specialized models.
Unified Multimodal Large Models (UMLMs) integrate understanding and generation capabilities within a single architecture. While this architectural unification, driven by the deep fusion of multimodal features, enhances model performance, it also introduces important yet underexplored safety challenges. Existing safety benchmarks predominantly focus on isolated understanding or generation tasks, failing to evaluate the holistic safety of UMLMs when handling diverse tasks under a unified framework. To address this, we introduce Uni-SafeBench, a comprehensive benchmark featuring a taxonomy of six major safety categories across seven task types. To ensure rigorous assessment, we develop Uni-Judger, a framework that effectively decouples contextual safety from intrinsic safety. Based on comprehensive evaluations across Uni-SafeBench, we uncover that while the unification process enhances model capabilities, it significantly degrades the inherent safety of the underlying LLM. Furthermore, open-source UMLMs exhibit much lower safety performance than multimodal large models specialized for either generation or understanding tasks. We open-source all resources to systematically expose these risks and foster safer AGI development.