analiza aiqbench
This research addresses critical limitations in current AIGC detection by utilizing a compact, multi-modal vision-language model that generalizes better to new generation techniques while providing interpretable explanations. The system's deployment on real-world social media platforms demonstrates feasibility in dynamic environments, yielding positive downstream effects on user engagement. For tech audiences, the key takeaway is the shift from single-modality approaches to robust, continuously curated pipelines capable of handling evolving generative threats.
Detecting AI-Generated Content on Social Media with Multi-modal Language Models
arXiv:2606.11200v1 Announce Type: new Abstract: Generative AI has enabled the creation of photorealistic images and videos that are increasingly disseminated on social media, often used for spam, misinformation, manipulation, and fraud. Existing AI-generated content (AIGC) detection methods face challenges including poor generalization to new generation models, reliance on single modalities, and lack of interpretable explanations. We present our pipeline that mitigates these issues by continuously curating diverse multi-modal social media data and training a compact vision-language model for detection and explanation. Our model achieves state-of-the-art detection performance on public benchmarks and demonstrates robust detection and explanation capabilities on internal social media datasets across multiple platforms. We deployed our model for post recommendation on social media platforms and observed positive downstream impacts on user engagement, demonstrating that it is feasible to perform effective AIGC detection in dynamic, real-world social media environments.
Abstract:Generative AI has enabled the creation of photorealistic images and videos that are increasingly disseminated on social media, often used for spam, misinformation, manipulation, and fraud. Existing AI-generated content (AIGC) detection methods face challenges including poor generalization to new generation models, reliance on single modalities, and lack of interpretable explanations. We present our pipeline that mitigates these issues by continuously curating diverse multi-modal social media data and training a compact vision-language model for detection and explanation. Our model achieves state-of-the-art detection performance on public benchmarks and demonstrates robust detection and explanation capabilities on internal social media datasets across multiple platforms. We deployed our model for post recommendation on social media platforms and observed positive downstream impacts on user engagement, demonstrating that it is feasible to perform effective AIGC detection in dynamic, real-world social media environments.
Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
arXiv:2606.11200 [cs.CL]
(or arXiv:2606.11200v1 [cs.CL] for this version)
Text integral sub licența deschisă a publicației · 153 cuvinte