ShieldGemma 2: Robust and Tractable Image Content Moderation

Authors:
  • WZWenjun Zeng
    Wenjun Zeng
  • DKDana Kurniawan
    Dana Kurniawan
  • RMRyan Mullins
    Ryan Mullins
  • YLYuchi Liu
    Yuchi Liu
  • TSTamoghna Saha
    Tamoghna Saha
  • DIDirichi Ike-Njoku
    Dirichi Ike-Njoku
  • JGJindong Gu
    Jindong Gu
  • YSYiwen Song
    Yiwen Song
  • CXCai Xu
    Cai Xu
  • JZJingjing Zhou
    Jingjing Zhou
  • AJAparna Joshi
    Aparna Joshi
  • SDShravan Dheep
    Shravan Dheep

Abstract
We introduce ShieldGemma 2, a 4B parameter image content moderation model built on Gemma 3. This model provides robust safety risk predictions across the following key harm categories: Sexually Explicit, Violence \& Gore, and Dangerous Content for synthetic images (e.g. output of any image generation model) and natural images (e.g. any image input to a Vision-Language Model). We evaluated on both internal and external benchmarks to demonstrate state-of-the-art performance compared to LlavaGuard \citep{helff2024llavaguard}, GPT-4o mini \citep{hurst2024gpt}, and the base Gemma 3 model \citep{gemma_2025} based on our policies. Additionally, we present a novel adversarial data generation pipeline which enables a controlled, diverse, and robust image generation. ShieldGemma 2 provides an open image moderation tool to advance multimodal safety and responsible AI development.