When Understanding Becomes a Risk: Authenticity and Safety Risks in the Emerging Image Generation Paradigm

TL;DR

This paper analyzes the safety risks of multimodal large language models (MLLMs), revealing they generate more unsafe and harder-to-detect images than diffusion models, raising concerns about their real-world safety implications.

Contribution

It systematically compares safety risks of MLLMs and diffusion models, highlighting their increased unsafe content generation and challenges in detection.

Findings

MLLMs produce more unsafe images than diffusion models.

MLLM-generated images are harder to detect even with retrained detectors.

Enhanced semantic understanding in MLLMs leads to greater safety risks.

Abstract

Recently, multimodal large language models (MLLMs) have emerged as a unified paradigm for language and image generation. Compared with diffusion models, MLLMs possess a much stronger capability for semantic understanding, enabling them to process more complex textual inputs and comprehend richer contextual meanings. However, this enhanced semantic ability may also introduce new and potentially greater safety risks. Taking diffusion models as a reference point, we systematically analyze and compare the safety risks of emerging MLLMs along two dimensions: unsafe content generation and fake image synthesis. Across multiple unsafe generation benchmark datasets, we observe that MLLMs tend to generate more unsafe images than diffusion models. This difference partly arises because diffusion models often fail to interpret abstract prompts, producing corrupted outputs, whereas MLLMs can comprehend these prompts and generate unsafe content. For current advanced fake image detectors, MLLM-generated images are also notably harder to identify. Even when detectors are retrained with MLLMs-specific data, they can still be bypassed by simply providing MLLMs with longer and more descriptive inputs. Our measurements indicate that the emerging safety risks of the cutting-edge generative paradigm, MLLMs, have not been sufficiently recognized, posing new challenges to real-world safety.