@inproceedings{zhang-etal-2026-aegis,
    title = "{AEGIS}: A Holistic Benchmark for Evaluating Forensic Analysis of {AI}-Generated Academic Images",
    author = "Zhang, Bo  and
      Ma, Tzu-Yen  and
      Tang, Zichen  and
      Ding, Junpeng  and
      Wang, Zirui  and
      Zhao, Yizhuo  and
      Gao, Peilin  and
      Xi, Zijie  and
      Ding, Zixin  and
      Sun, Haiyang  and
      Gao, Haocheng  and
      Liu, Yuan  and
      Wang, Liangjia  and
      Huang, Yiling  and
      Wang, Yujie  and
      Zhang, Yuyue  and
      Xi, Ronghui  and
      Li, Yuanze  and
      Liu, Jiacheng  and
      Yang, Zhongjun  and
      E, Haihong",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.acl-long.976/",
    pages = "21327--21363",
    ISBN = "979-8-89176-390-6",
    abstract = "We introduce AEGIS, A holistic benchmark for Evaluating forensic analysis of AI-Generated academic ImageS. Compared to existing benchmarks, AEGIS features three key advances: (1) Domain-Specific Complexity: covering seven academic categories with 39 fine-grained subtypes, exposing intrinsic forensic difficulty, where even GPT-5.1 reaches 48.80{\%} overall performance and expert models achieve only limited localization accuracy (IoU 30.09{\%}); (2) Diverse Forgery Simulations: modeling four prevalent academic forgery strategies across 25 generative models, with 11 yielding average forensic accuracy below 50{\%}, showing that forensics lag behind generative advances; and (3) Multi-Dimensional Forensic Evaluation: jointly assessing detection, reasoning, and localization, revealing complementary strengths between model families, with multimodal large language models (MLLMs) at 84.74{\%} accuracy in textual artifact recognition and expert detectors peaking at 79.54{\%} accuracy in binary authenticity detection. By evaluating 25 leading MLLMs, nine expert models, and one unified multimodal understanding and generation model, AEGIS serves as a diagnostic testbed exposing fundamental limitations in academic image forensics."
}