# FakeScope: Large Multimodal Expert Model for Transparent AI-Generated Image Forensics

Source: [FakeScope: Large Multimodal Expert Model for Transparent AI-Generated Image Forensics](<https://www.sota2.com/research/paper/68fa77af1e209131cea74b68>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-18T05:00:04.466Z

Authors: Yixuan Li, Yu Tian, Yipo Huang, Wei Lu, Shiqi Wang, Weisi Lin, Anderson Rocha

Published: 2025-03-31T16:12:48.000Z

## Abstract

The rapid and unrestrained advancement of generative artificial intelligence (AI) presents a double-edged sword. While enabling unprecedented creativity, it also facilitates the generation of highly convincing content, undermining societal trust. As image generation techniques become increasingly sophisticated, detecting synthetic images is no longer just a binary task--it necessitates explainable methodologies to enhance trustworthiness and transparency. However, existing detection models primarily focus on classification, offering limited explanatory insights. To address these limitations, we propose FakeScope, an expert large multimodal model (LMM) tailored for AI-generated image forensics, which not only identifies synthetic images with high accuracy but also delivers rich query-contingent forensic insights. At the foundation of our approach is FakeChain, a large-scale dataset containing structured forensic reasoning based on visual trace evidence, constructed via a human-machine collaborative framework. Then we develop FakeInstruct, the largest multimodal instruction tuning dataset to date, comprising two million visual instructions that instill nuanced forensic awareness into LMMs. Empowered by FakeInstruct, FakeScope achieves state-of-the-art performance in both closed-ended and open-ended forensic scenarios. It can accurately distinguish synthetic images, provide coherent explanations, discuss fine-grained forgery artifacts, and suggest actionable enhancement strategies. Notably, despite being trained exclusively on qualitative hard labels, FakeScope demonstrates remarkable zero-shot quantitative detection capability via our proposed token-based probability estimation strategy. Furthermore, it shows robust generalization across unseen image generators and performs reliably under in-the-wild scenarios.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Forensic Detection on FakeClass FakeBench (qualitative setting)](<https://www.sota2.com/research/sota/forensic-detection-on-fakeclass-fakebench-qualitative-setting.md>): Forensic Detection; FakeClass FakeBench (qualitative setting); Detection Rate (Fake): 99.5.

- [AIGC Detection on WildRF (Reddit)](<https://www.sota2.com/research/sota/aigc-detection-on-wildrf-reddit.md>): AIGC Detection; WildRF (Reddit); AP: 91.89.

- [AIGC Detection on WildRF (FB)](<https://www.sota2.com/research/sota/aigc-detection-on-wildrf-fb.md>): AIGC Detection; WildRF (FB); AP: 85.31.

- [AIGC Detection on WildRF (X)](<https://www.sota2.com/research/sota/aigc-detection-on-wildrf-x.md>): AIGC Detection; WildRF (X); AP: 92.96.

- [AIGC Detection on SynthWildX Firefly](<https://www.sota2.com/research/sota/aigc-detection-on-synthwildx-firefly.md>): AIGC Detection; SynthWildX Firefly; AP: 88.04.

- [AIGC Detection on WildRF and SynthWildX Combined Average](<https://www.sota2.com/research/sota/aigc-detection-on-wildrf-and-synthwildx-combined-average.md>): AIGC Detection; WildRF and SynthWildX Combined Average; mAP: 88.49.

- [AIGC Detection on SynthWildX Dalle3](<https://www.sota2.com/research/sota/aigc-detection-on-synthwildx-dalle3.md>): AIGC Detection; SynthWildX Dalle3; AP: 89.02.

- [AIGC Detection on SynthWildX (MJ)](<https://www.sota2.com/research/sota/aigc-detection-on-synthwildx-mj.md>): AIGC Detection; SynthWildX (MJ); AP: 83.72.

- [Forensic Attribution on FakeBench MCQ](<https://www.sota2.com/research/sota/forensic-attribution-on-fakebench-mcq.md>): Forensic Attribution; FakeBench MCQ; Yes/No Accuracy: 79.47.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2503.24267>)

[PDF](<https://arxiv.org/pdf/2503.24267.pdf>)
