# Unleashing the Potential of Vision-Language Models for Generalizable AI-Generated Image Detection

arXiv: https://arxiv.org/abs/2608.04935 · HTML: https://mlsanity.com/paper/2608.04935/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in MLLM jailbreaks and adversarial robustness, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 15 of its 22 references that are in the map.
- First posted on arXiv 5 Aug 2026.

## Placed in

- Multimodal LLMs / vision-language models / [MLLM jailbreaks and adversarial robustness](https://mlsanity.com/mllm/mllm-jailbreaks-and-adversarial-robustness/)
- Mechanistic interpretability / [Interpreting vision and multimodal models](https://mlsanity.com/mechinterp/interpreting-vision-and-multimodal-models/)

## Builds on (most cited first, 15 of 15)

- [LAION-5B: An open large-scale dataset for training next generation image-text models](https://mlsanity.com/paper/2210.08402/) (2022)
- [Reproducible scaling laws for contrastive language-image learning](https://mlsanity.com/paper/2212.07143/) (2022)
- [CosFace: Large Margin Cosine Loss for Deep Face Recognition](https://mlsanity.com/paper/1801.09414/) (2018)
- [SphereFace: Deep Hypersphere Embedding for Face Recognition](https://mlsanity.com/paper/1704.08063/) (2017)
- [Towards Universal Fake Image Detectors that Generalize Across Generative Models](https://mlsanity.com/paper/2302.10174/) (2023)
- [GenImage: A Million-Scale Benchmark for Detecting AI-Generated Image](https://mlsanity.com/paper/2306.08571/) (2023)
- [C2P-CLIP: Injecting Category Common Prompt in CLIP to Enhance Generalization in Deepfake Detection](https://mlsanity.com/paper/2408.09647/) (2024)
- [Dual Data Alignment Makes AI-Generated Image Detector Easier Generalizable](https://mlsanity.com/paper/2505.14359/) (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)
- [Breaking Latent Prior Bias in Detectors for Generalizable AIGC Image Detection](https://mlsanity.com/paper/2506.00874/) (2025)
- [Brought a Gun to a Knife Fight: Modern VFM Baselines Outgun Specialized Detectors on In-the-Wild AI Image Detection](https://mlsanity.com/paper/2509.12995/) (2025)
- [Rethinking Cross-Generator Image Forgery Detection through DINOv3](https://mlsanity.com/paper/2511.22471/) (2025)
- [Simplicity Prevails: The Emergence of Generalizable AIGI Detection in Visual Foundation Models](https://mlsanity.com/paper/2602.01738/) (2026)
- [SSAFE: Simple and Strong AI-Generated Image Detection via Frozen Vision Encoders](https://mlsanity.com/paper/2606.08634/) (2026)

## Abstract (arXiv)

Recent work has shown that a simple linear probe on frozen representations from modern vision foundation models (VFMs) can achieve state-of-the-art AIGI detection performance, substantially outperforming specialized detectors in challenging in-the-wild scenarios. This finding has established DINOv3 as the dominant foundation-model baseline for subsequent improvements. However, we find that the vision-language model Perception Encoder (PE) holds greater potential for AIGI detection, because its language-aligned representation preserves high-level provenance semantics. Specifically, PE exhibits stronger local provenance organization than DINOv3 in its frozen feature space. However, semantic-agnostic linear probing fails to exploit this structure, as PE-Linear still underperforms DINOv3-Linear by 4.1% on In-the-Wild. Based on this observation, we propose Semantic Prototype Calibration (SPC), which constructs category prototypes from forensic semantic information and calibrates them with supervised data. We apply SPC to PE and refer to the resulting detector as PE-SPC. Our analysis shows that this simple design achieves stronger generalization. Across cross-generator, post-processing, and in-the-wild benchmarks, PE-SPC surpasses the previous DINOv3 baseline and achieves new state-of-the-art results.