# Seeing Before Reasoning: A Unified Framework for Generalizable and Explainable Fake Image Detection

arXiv: https://arxiv.org/abs/2509.25502 · HTML: https://mlsanity.com/paper/2509.25502/

## Key facts

- Cited by 20 papers in the map (data through 14 Oct 2026).
- Placed in Multimodal reasoning, one of 17 directions in LLM reasoning.
- Builds on 26 of its 65 references that are in the map.
- First posted on arXiv 29 Sep 2025.

## Placed in

- LLM reasoning / [Multimodal reasoning](https://mlsanity.com/lr/multimodal-reasoning/)
- Multimodal LLMs / vision-language models / [MLLM jailbreaks and adversarial robustness](https://mlsanity.com/mllm/mllm-jailbreaks-and-adversarial-robustness/)

## Builds on (most cited first, 26 of 26)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [BLINK: Multimodal Large Language Models Can See but Not Perceive](https://mlsanity.com/paper/2404.12390/) (2024)
- [SWIFT:A Scalable lightWeight Infrastructure for Fine-Tuning](https://mlsanity.com/paper/2408.05517/) (2024)
- [A Sanity Check for AI-generated Image Detection](https://mlsanity.com/paper/2406.19435/) (2024)
- [Orthogonal Subspace Decomposition for Generalizable AI-Generated Image Detection](https://mlsanity.com/paper/2411.15633/) (2024)
- [Echo-4o: Harnessing the Power of GPT-4o Synthetic Images for Improved Image Generation](https://mlsanity.com/paper/2508.09987/) (2025)
- [Improving Synthetic Image Detection Towards Generalization: An Image Transformation Perspective](https://mlsanity.com/paper/2408.06741/) (2024)
- [MME-Survey: A Comprehensive Survey on Evaluation of Multimodal LLMs](https://mlsanity.com/paper/2411.15296/) (2024)
- [Spot the Fake: Large Multimodal Model-Based Synthetic Image Detection with Artifact Explanation](https://mlsanity.com/paper/2503.14905/) (2025)
- [LEGION: Learning to Ground and Explain for Synthetic Image Detection](https://mlsanity.com/paper/2503.15264/) (2025)
- [Dual Data Alignment Makes AI-Generated Image Detector Easier Generalizable](https://mlsanity.com/paper/2505.14359/) (2025)
- [GPT-ImgEval: A Comprehensive Benchmark for Diagnosing GPT4o in Image Generation](https://mlsanity.com/paper/2504.02782/) (2025)
- [Real-Time Deepfake Detection in the Real-World](https://mlsanity.com/paper/2406.09398/) (2024)
- [Is Artificial Intelligence Generated Image Detection a Solved Problem?](https://mlsanity.com/paper/2505.12335/) (2025)
- [Learning or Self-aligning? Rethinking Instruction Fine-tuning](https://mlsanity.com/paper/2402.18243/) (2024)
- [On the generalization of language models from in-context learning and finetuning: a controlled study](https://mlsanity.com/paper/2505.00661/) (2025)
- [X2-DFD: A framework for eXplainable and eXtendable Deepfake Detection](https://mlsanity.com/paper/2410.06126/) (2024)
- [Ivy-Fake: A Unified Explainable Framework and Benchmark for Image and Video AIGC Detection](https://mlsanity.com/paper/2506.00979/) (2025)
- [AIGI-Holmes: Towards Explainable and Generalizable AI-Generated Image Detection via Multimodal Large Language Models](https://mlsanity.com/paper/2507.02664/) (2025)
- [Toward Generalizable Forgery Detection and Reasoning](https://mlsanity.com/paper/2503.21210/) (2025)
- [Breaking Latent Prior Bias in Detectors for Generalizable AIGC Image Detection](https://mlsanity.com/paper/2506.00874/) (2025)
- [MLLM-Enhanced Face Forgery Detection: A Vision-Language Fusion Solution](https://mlsanity.com/paper/2505.02013/) (2025)
- [Unified Multimodal Models as Auto-Encoders](https://mlsanity.com/paper/2509.09666/) (2025)
- [VLForgery Face Triad: Detection, Localization and Attribution via Multimodal Large Language Models](https://mlsanity.com/paper/2503.06142/) (2025)
- [Guard Me If You Know Me: Protecting Specific Face-Identity from Deepfakes](https://mlsanity.com/paper/2505.19582/) (2025)

## Built on by (most cited first, 20 of 20)

- [RealGen: Photorealistic Text-to-Image Generation via Detector-Guided Rewards](https://mlsanity.com/paper/2512.00473/) (2025)
- [Venus-DeFakerOne: Unified Fake Image Detection & Localization](https://mlsanity.com/paper/2605.14091/) (2026)
- [DINO-Detect: A Simple yet Effective Framework for Blur-Robust AI-Generated Image Detection](https://mlsanity.com/paper/2511.12511/) (2025)
- [DGS-Net: Distillation-Guided Gradient Surgery for CLIP Fine-Tuning in AI-Generated Image Detection](https://mlsanity.com/paper/2511.13108/) (2025)
- [AlignGemini: Generalizable AI-Generated Image Detection Through Task-Model Alignment](https://mlsanity.com/paper/2512.06746/) (2025)
- [FakeVLM-R1: Internalizing Physical Laws via CoT for Synthetic Image Detection](https://mlsanity.com/paper/2605.30062/) (2026)
- [AgentFoX: LLM Agent-Guided Fusion with eXplainability for AI-Generated Image Detection](https://mlsanity.com/paper/2603.23115/) (2026)
- [ReAlign: Generalizable Image Forgery Detection via Reasoning-Aligned Representation](https://mlsanity.com/paper/2605.16080/) (2026)
- [Deep Residual Injection for Full-Spectrum Forensic Signal Perception in Multimodal Large Language Models](https://mlsanity.com/paper/2606.15880/) (2026)
- [ForgeryVCR: Visual-Centric Reasoning via Efficient Forensic Tools in MLLMs for Image Forgery Detection and Localization](https://mlsanity.com/paper/2602.14098/) (2026)
- [A Difference-in-Difference Approach to Detecting AI-Generated Images](https://mlsanity.com/paper/2602.23732/) (2026)
- [Where, What, Why, and Importance: Structured Defect Grounding for Text-to-Image Feedback](https://mlsanity.com/paper/2606.06113/) (2026)
- [LaP-Forensics: Latent-Pixel Consistency Guided Multimodal Reasoning for Deepfake Detection](https://mlsanity.com/paper/2607.25962/) (2026)
- [Evidence-Grounded Forensic Reasoning for Detecting and Grounding Multi-Modal Media Manipulation](https://mlsanity.com/paper/2608.08009/) (2026)
- [Anchor-Regularized Adaptation for Generalizable AI-Generated Image Detection with DINOv3](https://mlsanity.com/paper/2608.15196/) (2026)
- [NormGuard: Reward-Preserving Norm Constraints in Flow-Matching Reinforcement Learning](https://mlsanity.com/paper/2606.27771/) (2026)
- [Temporal Concentration from Rollout Errors: Implicit Preference Optimization for Text-to-Video Diffusion](https://mlsanity.com/paper/2607.28058/) (2026)
- [Editable Visual Design](https://mlsanity.com/paper/2609.04034/) (2026)
- [ManiVid: Unified and Explainable Forensic Analysis of Manipulated Videos](https://mlsanity.com/paper/2609.30934/) (2026)
- [From Sharp Eyes to Expert Mind: Internalizing Expert Knowledge in MLLMs for Tampered Text Detection](https://mlsanity.com/paper/2609.36145/) (2026)

## Abstract (arXiv)

Detecting AI-generated images with multimodal large language models (MLLMs) has gained increasing attention, due to their rich world knowledge, common-sense reasoning, and potential for explainability. However, naively applying those MLLMs for detection often leads to suboptimal performance. We argue that the root of this failure lies in a fundamental mismatch: MLLMs are asked to reason about fakes before they can truly see them. First, they do not really see: existing MLLMs' vision encoders are primarily optimized for semantic-oriented recognition rather than the perception of low-level signals, leaving them insensitive to subtle forgery traces. Without access to reliable perceptual evidence, the model grounds its judgment on incomplete and limited visual observations. Second, existing finetuning data for detection typically uses narrow, instruction-style formats, which diverge sharply from the diverse, heterogeneous distributions seen in pretraining. In the absence of meaningful visual cues, the model therefore exploits these linguistic shortcuts, resulting in catastrophic forgetting of pretrained knowledge (even the basic dialogue capabilities). In response, we advocate for a new paradigm: seeing before reasoning. We propose that MLLMs should first be trained to perceive artifacts-strengthening their artifact-aware visual perception-so that subsequent reasoning is grounded in actual observations. We therefore propose Forensic-Chat, a generalizable, explainable, and still-conversational (for multi-round dialogue) assistant for fake image detection. We also propose ExplainFake-Bench, a benchmark tailored for the evaluation of the MLLM's explainability for image forensics from five key aspects. Extensive experiments show its superiority of generalization and genuinely reliable explainability.