Metraipolis › archive › dc_9e9fa0940d5b
Beyond Generation: Using Multimodal LLMs as Detectors of LLM-Generated Deepfake Facial Images — arXiv:2610.21760
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.21760
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T10:35:28.507Z
- SHA-256 of the archived text
8fffb42761269c9248eeed126270262e2700c60001fa1d0f401bf48f3dd79c0a- Stable URL of this page
https://metraipolis.com/documento/dc_9e9fa0940d5b
How to cite
Beyond Generation: Using Multimodal LLMs as Detectors of LLM-Generated Deepfake Facial Images — arXiv:2610.21760 — arXiv (en), https://arxiv.org/abs/2610.21760 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_9e9fa0940d5b (Metraipolis archive dc_9e9fa0940d5b, sha256 8fffb42761269c9248eeed126270262e2700c60001fa1d0f401bf48f3dd79c0a)
BibTeX
@misc{dc9e9fa0940d5b,
title = {Beyond Generation: Using Multimodal LLMs as Detectors of LLM-Generated Deepfake Facial Images — arXiv:2610.21760},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.21760},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_9e9fa0940d5b; sha256 8fffb42761269c9248eeed126270262e2700c60001fa1d0f401bf48f3dd79c0a},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.21760. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
As image generation models become increasingly capable of producing highly realistic content, traditional forensic techniques are struggling to keep up. This paper investigates whether MLLMs can instead be used as a form of forensic analysis by identifying synthetic media. Specifically, the study aims to determine whether these models detect fake images by recognising physical inconsistencies, or whether they rely on digital artefacts such as embedded this http URL MLLMs: GPT-4o, GPT-5, GPT-5.2, Gemini 2.5, Gemini 3, and Seed 1.6, are benchmarked using a dataset of 400 images. Fifty images are authentic photographs, while the remaining 350 are generated using seven leading image models: Gemini 3, Gemini 2.5, Flux.2 Pro, GPT-Image-1, GPT-Image-1.5, Seedream 4.0, and Seedream 4.5. The evaluation is structured in three stages: single-image classification, pairwise comparison, and a focused "self-detection" experiment on Gemini. Several MLLMs achieved a high detection accuracy, with GPT-5.2 reaching a peak of 97.9% in the pairwise experiment, against specific generators and an overall mean pairwise accuracy of 82.5% by identifying structural or physical inconsistencies within images. However, results from the self-detection experiment indicate that the presence of invisible SynthID watermarks leads to a 10% reduction in Gemini's detection accuracy compared to the clean control group. One possible explanation is that SynthID embedding reduces some of the visible artefacts the models rely on for this http URL results imply that MLLMs can detect structural artefacts without relying on watermarks, but performance varies substantially across model generations. The findings from comparison C5 also suggest that SynthID embedding may suppress some of the structural artefacts used during detection. Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Emerging Technologies (cs.ET). Pagina del lavoro: https://arxiv.org/abs/2610.21760 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-16T08:52:02.693Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).