Skip to content
NewA calmer, faster ZeroHour

EditSafetyBench finds image editors miss unsafe images with neutral prompts

Original title: Text-Aware, Vision-Blind: How Multimodal Image Editing Systems Handle Unsafe Inputs

arXiv cs.CRby Ziqing Yang ()1 min read
AI summary · x-ai/grok-4.7Summaries by Vani AI

Gemini and GPT refuse most unsafe-intent prompts but only about a quarter of unsafe images with neutral instructions.

The paper introduces EditSafetyBench, with 5,940 image-instruction pairs across 11 risk categories and three edit intents, and probes Gemini, GPT, Qwen API, and Qwen Local. Gemini and GPT refuse 87.1% and 81.6% of unsafe-intent prompts, versus 24.2% and 22.4% when unsafe images accompany neutral instructions. Sexual content shows the strongest visual sensitivity, while political-content decisions diverge sharply between GPT and Gemini.

Key facts
  • The benchmark contains 5,940 image-instruction pairs.
  • Gemini refuses 24.2% of unsafe-image, neutral-instruction cases.
  • GPT allows 70.4% of unsafe political edits; Gemini allows 34.8%.
  • Qwen API and a local Qwen system were also probed.
Explain this AI news with Vani AIWhat was announced, what is new and why it matters, in a new chat

From the original article

1 min read at arxiv.org

Multimodal image generation systems have made rapid progress and increasingly emphasize content safety. However, their safety mechanisms tend to rely on a single modality and lack joint consideration of both visual content and text prompts, particularly in instruction-guided image editing, where the uploaded image may itself carry risk.

arXiv cs.CR

By Ziqing Yang · arxiv.org

Read the full article at arXiv cs.CR

A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.