Elamur Research⌕
☰
home / papers / 2026 / Detecting AI-Generated Content on Social Media with Multi-modal Language Models

Detecting AI-Generated Content on Social Media with Multi-modal Language Models

Paper · 2026 · Chenyang Yang, Shen Yan, Yibo Yang et al.

Summary

The paper tackles automatic recognition of AI-generated images and videos in social media posts. It addresses three weaknesses of earlier AIGC detectors: they degrade on content from newer generators, they use only one modality, and they rarely justify their decisions. The authors, from Meta, built a data pipeline that keeps harvesting labelled posts from Facebook, Instagram and Threads. Labels come from keyword heuristics on post text and comments, refined by embedding clustering and human checks. On this data they train IFM-AIGCSpotter-3B, a compact LLaVA-style vision-language model. Training runs in three stages: captioning, multi-task tuning and AIGC detection with explanations. The model reaches high precision and recall on internal data from all three platforms. Without any adaptation, it also beats ten published detectors on the public Chameleon benchmark by more than 20 points. In production, the system demotes creators who post heavily AI-generated content, and online A/B tests showed small but statistically significant engagement gains. This shows that multimodal, continuously retrained detection can work at real-world scale.

Key points

  • Architecture: Llama3.2-3B-Instruct serves as the language model, paired with a 300M-parameter Perception Encoder. Images are cropped to 448×448 and pooled to 256 visual tokens. A single linear projection connects the two, initialised with zero weights.
  • Training data: 60M+ caption pairs from Shutterstock, social posts and 164K human-annotated images. Multi-task tuning used 130K samples covering NER, hashtags, keywords and genre. AIGC tuning used 200K samples (30K positive, 170K negative), with explanations generated by Llama4-Maverik.
  • Internal results: Overall precision is 0.886 and recall is 0.853. Threads scores highest, at 0.926 precision and 0.919 recall.
  • Ablations: Randomly dropping post text during training is key for recall, because it stops the model from leaning too heavily on textual cues. Multi-frame video input brings significant gains, and Perception Encoder gives a slight precision edge over CLIP.
  • Benchmark and deployment: The model scores 86.0% accuracy on Chameleon, against 62.60 for the best baseline (AIDE). It runs on 4 H100 nodes and screens every post above 500 views. After six weeks, marginal users saw +0.21% views and global Young-Adult sessions rose +0.22%.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Carnegie Mellon UniversityPittsburgh · United States598 works · 191,226 citations

 Fields, topics and concepts

📚 Shelves 7The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksArts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 worksComputer Vision and Pattern Recognition · 3,486 worksSocial Sciences · 1,954 worksSociology and Political Science · 626 works

🧭 Topics 2Research subjects and languages the work is about.Language models & pretraining · 3,894 worksVision & language · 811 works

Cite

Cite as Chenyang Yang, Shen Yan, Yibo Yang, Litao Hu, Yuchen Liu, Yuan Zeng, Hanchao Yu, Yinan Zhu, Sumedha Singla, Brian Vanover, Huijun Qian, Zihao Wang, Fujun Liu, Aashu Singh, Jianyu Wang and Xuewen Zhang. 2026. Detecting AI-Generated Content on Social Media with Multi-modal Language Models. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 6: Industry Track).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.acl-industry.15

@inproceedings{yang2026detecting,
  title = {{Detecting AI-Generated Content on Social Media with Multi-modal Language Models}},
  author = {Chenyang Yang, Shen Yan, Yibo Yang, Litao Hu, Yuchen Liu, Yuan Zeng, Hanchao Yu, Yinan Zhu, Sumedha Singla, Brian Vanover, Huijun Qian, Zihao Wang, Fujun Liu, Aashu Singh, Jianyu Wang and Xuewen Zhang},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 6: Industry Track)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-industry.15},
  url = {https://doi.org/10.18653/v1/2026.acl-industry.15}
}

References

28 references, 7 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Associated Press. 2024. Trump arrested? putin jailed? fake ai images spread online. find
  2. Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, and 1 others. 2025. Qwen2. 5-vl technical report. arXiv preprint arXiv:2502.13923. — Qwen2.5-VL Technical Report
  3. Daniel Bolya, Po-Yao Huang, Peize Sun, Jang Hyun Cho, Andrea Madotto, Chen Wei, Tengyu Ma, Jiale Zhi, Jathushan Rajasegaran, Hanoona Rasheed, and 1 others. 2025. Perception encoder: The best visual embeddings are not at the output of the network. arXiv preprint arXiv:2504.13181. — [arXiv:2504.13181](https://arxiv.org/abs/2504.13181) find
  4. Tim Brooks, Bill Peebles, Connor Holmes, Will DePue, Yufei Guo, Li Jing, David Schnurr, Joe Taylor, Troy Luhman, Eric Luhman, Clarence Ng, Ricky Wang, and Aditya Ramesh. 2024. Video generation models as world simulators. find
  5. Nuria Alina Chandra, Ryan Murtfeldt, Lin Qiu, Arnab Karmakar, Hannah Lee, Emmanuel Tanumihardja, Kevin Farhat, Ben Caffee, Sejin Paik, Changyeon Lee, and 1 others. 2025. Deepfake-eval-2024: A multi-modal in-the-wild benchmark of deepfakes circulated in 2024. arXiv preprint arXiv:2503.02857. find
  6. Zhe Chen, Weiyun Wang, Yue Cao, Yangzhou Liu, Zhangwei Gao, Erfei Cui, Jinguo Zhu, Shenglong Ye, Hao Tian, Zhaoyang Liu, and 1 others. 2024a. Expanding performance boundaries of open-source multimodal models with model, data, and test-time scaling. arXiv preprint arXiv:2412.05271.
  7. Zhe Chen, Jiannan Wu, Wenhai Wang, Weijie Su, Guo Chen, Sen Xing, Muyan Zhong, Qinglong Zhang, Xizhou Zhu, Lewei Lu, and 1 others. 2024b. Internvl: Scaling up vision foundation models and aligning Davide Cozzolino, Giovanni Poggi, Riccardo Corvi, Matthias Nießner, and Luisa Verdoliva. 2024. Raising… find
  8. Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Alex Vaughan, and 1 others. 2024. The llama 3 herd of models. arXiv preprint arXiv:2407.21783.
  9. Runsheng Huang, Liam Dugan, Yue Yang, and Chris Callison-Burch. 2024. Miragenews: Multimodal realistic ai-generated news detection. arXiv preprint arXiv:2410.09045. find
  10. Shan Jia, Reilin Lyu, Kangran Zhao, Yize Chen, Zhiyuan Yan, Yan Ju, Chuanbo Hu, Xin Li, Baoyuan Wu, and Siwei Lyu. 2024. Can chatgpt detect deepfakes? a study of using multimodal large language models for media forensics. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recog… find
  11. Mamadou Keita, Wassim Hamidouche, Hessen Bougu-effa Eutamene, Abdelmalik Taleb-Ahmed, and Abdenour Hadid. 2024. Fidavl: Fake image detection and attribution using vision-language model. In International Conference on Pattern Recognition, pages 160–176. Springer. find
  12. Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. 2023. Visual instruction tuning. Advances in neural information processing systems, 36:34892– 34916.
  13. Jiawei Liu, Fanrui Zhang, Jiaying Zhu, Esther Sun, Qiang Zhang, and Zheng-Jun Zha. 2024. Forgerygpt: Multimodal large language model for explainable image forgery detection and localization. arXiv preprint arXiv:2410.10238. — [arXiv:2410.10238](https://arxiv.org/abs/2410.10238) find
  14. Scott McCloskey and Michael Albright. 2018. Detecting gan-generated imagery using color cues. arXiv preprint arXiv:1812.08247. — [arXiv:1812.08247](https://arxiv.org/abs/1812.08247) find
  15. Utkarsh Ojha, Yuheng Li, and Yong Jae Lee. 2023. Towards universal fake image detectors that generalize across generative models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 24480–24489. find
  16. Md Awsafur Rahman, Bishmoy Paul, Najibul Haque Sarker, Zaber Ibn Abdul Hakim, and Shaikh Anowarul Fattah. 2023. Artifact: A large-scale dataset with artificial and factual images for generalizable and robust synthetic image detection. In 2023 IEEE International Conference on Image Processing (ICIP),… find
  17. Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Björn Ommer. 2022. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684–10695.
  18. Zhiyu Wu, Xiaokang Chen, Zizheng Pan, Xingchao Liu, Wen Liu, Damai Dai, Huazuo Gao, Yiyang Ma, Chengyue Wu, Bingxuan Wang, and 1 others. 2024. Deepseek-vl2: Mixture-of-experts vision-language models for advanced multimodal understanding. arXiv preprint arXiv:2412.10302.
  19. Shilin Yan, Ouxiang Li, Jiayin Cai, Yanbin Hao, Xiaolong Jiang, Yao Hu, and Weidi Xie. 2024. A sanity check for ai-generated image detection. arXiv preprint arXiv:2406.19435. — [arXiv:2406.19435](https://arxiv.org/abs/2406.19435) find
  20. Junyan Ye, Baichuan Zhou, Zilong Huang, Junan Zhang, Tianyi Bai, Hengrui Kang, Jun He, Honglin Lin, Zihao Wang, Tong Wu, and 1 others. 2024. Loki: A comprehensive synthetic data detection benchmark using large multimodal models. arXiv preprint arXiv:2410.09732. — [arXiv:2410.09732](https://arxiv.org/abs/2410.09732) find
  21. Andreas Rossler, Davide Cozzolino, Luisa Verdoliva, Christian Riess, Justus Thies, and Matthias Nießner. 2019. Faceforensics++: Learning to detect manipulated facial images. In Proceedings of the IEEE/CVF international conference on computer vision, pages 1–11. find
  22. Christian Tarsney. 2025. Deception and manipulation in generative ai. Philosophical Studies, pages 1–23. find
  23. Francesco Tassone, Luca Maiano, and Irene Amerini. 2024. Continuous fake media detection: adapting deepfake detectors to new generative techniques. Computer Vision and Image Understanding, 249:104143. find
  24. Peng Wang, Shuai Bai, Sinan Tan, Shijie Wang, Zhihao Fan, Jinze Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, and 1 others. 2024. Qwen2-vl: Enhancing vision-language model’s perception of the world at any resolution. arXiv preprint arXiv:2409.12191.
  25. Sheng-Yu Wang, Oliver Wang, Richard Zhang, Andrew Owens, and Alexei A Efros. 2020. Cnn-generated images are surprisingly easy to spot... for now. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8695–8704. find
  26. Haiquan Wen, Tianxiao Li, Zhenglin Huang, Yiwei He, and Guangliang Cheng. 2025a. Busterx++: Towards unified cross-modal ai-generated content detection and explanation with mllm. arXiv preprint arXiv:2507.14632. — [arXiv:2507.14632](https://arxiv.org/abs/2507.14632) find
  27. Siwei Wen, Junyan Ye, Peilin Feng, Hengrui Kang, Zichen Wen, Yize Chen, Jiang Wu, Wenjun Wu, Conghui He, and Weijia Li. 2025b. Spot the fake: Large multimodal model-based synthetic image detection with artifact explanation. Preprint, arXiv:2503.14905. — [arXiv:2503.14905](https://arxiv.org/abs/2503.14905) find
  28. Wired. 2024. Deepfake scams are distorting reality itself. https://www.wired.com/story/ youre-not-ready-for-ai-powered-scams/. ↗

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/yang-2026-detecting-ai-generated-content-social/