Defake-o3: From Speculative Rationales to Verifiable Evidence for Explainable AIGI Detection

2026-08-17

AI Summary

This study addresses the limitations of ambiguous explanations, hallucinations, and insufficient visual evidence in existing AI-generated image detectors by proposing Defake-o3, an interpretable detection framework. Integrating interactive visual search with an evidence verifier, this method employs reinforcement learning rewards to guide multimodal large language models in aligning speculative reasoning with verifiable evidence. Additionally, the GroundFake dataset and FakeFrontier benchmark are introduced to facilitate evaluation. Experimental results demonstrate that Defake-o3 significantly improves both detection accuracy and explanation quality across multiple benchmarks. By generating precise, verifiable, and persuasive evidence, the proposed framework effectively enhances the overall trustworthiness of AI-generated image detection systems.

Abstract

The rapid progress of image generation models calls for AI-generated image (AIGI) detectors that are not only accurate but also explainable and reliable. While MLLM-based detectors can provide natural language explanations, existing methods often generate speculative rationales: they rely on vague or hallucinated artifacts, miss subtle localized flaws from the latest generators, and fail to provide evidence that can be visually verified. We present Defake-o3, an explainable AIGI detector that moves from speculative rationales to verifiable evidence. It combines interactive visual search with verifier-guided evidence alignment: the model iteratively zooms into suspicious regions to inspect fine-grained details, while an Evidence Verifier, trained from human verification annotations, provides reinforcement learning rewards that favor grounded evidence and penalize baseless claims. To support this objective, we construct GroundFake, a dataset designed for grounded explainable detection, with localized bounding-box evidence, human verification based on visual grounding and artifact specificity, corrected reasoning trajectories, and valid/invalid evidence supervision. We further introduce FakeFrontier, an out-of-distribution benchmark built from real images and outputs of 10 recent generators, together with an MLLM-based protocol for evaluating evidence quality and persuasiveness. Experiments on GroundFake, FakeFrontier, and additional out-of-distribution benchmarks show that Defake-o3 improves both detection accuracy and explanation quality, producing more localized, verifiable, and persuasive evidence.

Problem

• Explainable AIGI Detection
• Speculative Rationales
• Verifiable Evidence
• Hallucinated Artifacts
• MLLM-based Detectors

Innovation

• Verifiable Evidence
• Evidence Verifier
• Interactive Visual Search
• GroundFake Dataset
• Reinforcement Learning