Prefilled responses enhance zero-shot detection of AI-generated images
This addresses the need for reliable, zero-shot detection of AI-generated images to prevent misuse, offering a practical solution for security and verification domains, though it is incremental as it builds on existing VLMs.
The paper tackled the problem of detecting AI-generated images without training data by using pre-trained Vision-Language Models (VLMs) with a simple prefilling technique, improving Macro F1 scores by up to 24% on diverse benchmarks.
As AI models generate increasingly realistic images, growing concerns over potential misuse underscore the need for reliable detection. Traditional supervised detection methods depend on large, curated datasets for training and often fail to generalize to novel, out-of-domain image generators. As an alternative, we explore pre-trained Vision-Language Models (VLMs) for zero-shot detection of AI-generated images. We evaluate VLM performance on three diverse benchmarks encompassing synthetic images of human faces, objects, and animals produced by 16 different state-of-the-art image generators. While off-the-shelf VLMs perform poorly on these datasets, we find that their reasoning can be guided effectively through simple response prefilling -- a method we call Prefill-Guided Thinking (PGT). In particular, prefilling a VLM response with the task-aligned phrase "Let's examine the style and the synthesis artifacts" improves the Macro F1 scores of three widely used open-source VLMs by up to 24%.