Where's the liability in the Generative Era? Recovery-based Black-Box Detection of AI-Generated Content

2025-05-02

AI Summary

The proliferation of generative AI image misuse necessitates efficient, deployable detection methods. This paper proposes a black-box detection framework that requires neither model weights nor access to authentic training images. Its core is a “mask-reconstruct” discriminative paradigm: it identifies whether an image was generated by a target diffusion model by measuring fidelity discrepancies in its reconstruction after local masking. To accommodate uncontrolled commercial APIs, we introduce lightweight surrogate model distillation and an API interaction mechanism. Furthermore, a diffusion-model distribution alignment strategy enhances cross-model generalization. Evaluated on eight mainstream diffusion model variants, our method achieves a 4.31% average precision (mAP) improvement over state-of-the-art baselines. The framework demonstrates superior practical deployability and robustness under real-world constraints, including limited computational resources and opaque model interfaces.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Generation

Application Category

• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

The recent proliferation of photorealistic images created by generative models has sparked both excitement and concern, as these images are increasingly indistinguishable from real ones to the human eye. While offering new creative and commercial possibilities, the potential for misuse, such as in misinformation and fraud, highlights the need for effective detection methods. Current detection approaches often rely on access to model weights or require extensive collections of real image datasets, limiting their scalability and practical application in real world scenarios. In this work, we introduce a novel black box detection framework that requires only API access, sidestepping the need for model weights or large auxiliary datasets. Our approach leverages a corrupt and recover strategy: by masking part of an image and assessing the model ability to reconstruct it, we measure the likelihood that the image was generated by the model itself. For black-box models that do not support masked image inputs, we incorporate a cost efficient surrogate model trained to align with the target model distribution, enhancing detection capability. Our framework demonstrates strong performance, outperforming baseline methods by 4.31% in mean average precision across eight diffusion model variant datasets.

Problem

• Detect AI-generated images without model weights
• Overcome scalability limits of current detection methods
• Address misuse risks of photorealistic generative content

Innovation

• Black-box detection via API access only
• Corrupt and recover strategy for reconstruction assessment
• Cost-efficient surrogate model for alignment enhancement