Elamur Research⌕
☰
home / papers / 2026 / Multimodal Analysis of State-Funded News Coverage of the Israel–Hamas War on YouTube Shorts

Multimodal Analysis of State-Funded News Coverage of the Israel–Hamas War on YouTube Shorts

Paper · 2026 · Daniel Miehling and Sandra Kübler

Summary

Short vertical videos are now a major way people get news on YouTube, but little research has examined how geopolitical conflicts appear in this format, especially from state-funded broadcasters. The authors build a multimodal pipeline to study Israel–Hamas war coverage in YouTube Shorts. The pipeline has four parts:

  • Whisper transcription.
  • Dependency-based linking of conflict-related aspect terms.
  • Aspect-based sentiment analysis (ABSA), using a model fine-tuned on an extended gold standard.
  • Visual scene classification with a vision–language model and a new seven-category taxonomy.

They collected Shorts for one year from Al Jazeera, BBC, Deutsche Welle and TRT World, then examined how the distributions of sentiment and visual scenes differed between broadcasters and shifted over the collection period. Sentiment toward specific actors varies by broadcaster and period. The visual scene labels track real events on the ground. A notable methodological finding is that a mid-sized domain-adapted transformer beat larger models and a fine-tuned LLM on ABSA. This suggests resource-efficient tools are adequate for computational social science and humanities work. The pipeline is presented as reusable for TikTok and Instagram.

Key points

  • Dataset: 2,371 Shorts collected via the YouTube API, starting two weeks after October 7, 2023. After removing 36 mostly non-English videos, 2,335 remain. The study also analyses more than 94,000 visual frames.
  • Transcription and aspect linking:
    • Transcripts come from Whisper large-v3.
    • Aspects are linked with the SuPar biaffine-dep-en parser, trained on English UD treebanks.
    • A 56-form lexicon in ten categories (e.g., Israel, Palestine, Islamism, Netanyahu, Sinwar) yields 3,252 aspect instances across 782 videos.
  • Training data: The ABSA gold standard of Miehling et al. (2025) was extended. The authors added 1,255 manually validated predictions from DeBERTa-v3-large-absa-v1.1, keeping only those with confidence ≥ 0.75.
  • Model comparison (F1):
ModelF1
DeBERTa-v3-base81.89
DeBERTa-v3-large-absa-v1.179.65
DeBERTa-v3-large78.43
RoBERTa-base76.32
Qwen2.5-7B (QLoRA)72.50
  • Scene taxonomy: Developed with Qwen3-VL (4B) by narrowing 14 categories to 12 and then 7 over about 4,000 frames. The final types include:
    • combat
    • destruction/humanitarian crisis
    • political events
    • news/interview settings
    • protest
    • symbolic/religious ritual
    • other/unknown

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

🎓Indiana UniversityBloomington · United States41 works · 3,663 citations

 Fields, topics and concepts

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 worksSocial Sciences · 1,772 worksGeneral Social Sciences · 227 works

🧭 Topics 1Research subjects and languages the work is about.Vision & language · 657 works

Cite

Cite as Daniel Miehling and Sandra Kübler. 2026. Multimodal Analysis of State-Funded News Coverage of the Israel–Hamas War on YouTube Shorts. Proceedings of the 3rd Workshop on Natural Language Processing for Political Sciences (PoliticalNLP 2026).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.63317/5kxkswtz4mez

@inproceedings{miehling2026multimodal,
  title = {{Multimodal Analysis of State-Funded News Coverage of the Israel–Hamas War on YouTube Shorts}},
  author = {Daniel Miehling and Sandra Kübler},
  booktitle = {Proceedings of the 3rd Workshop on Natural Language Processing for Political Sciences (PoliticalNLP 2026)},
  year = {2026},
  doi = {10.63317/5kxkswtz4mez},
  url = {https://doi.org/10.63317/5kxkswtz4mez}
}

References

25 references, 5 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Precious N. Chatterje-Doody and Rhys Crilley. 2019. Making sense of emotions and affective investments in war: Rt and the syrian conflict on youtube. Media and Communication, 7(3):167– 178. find
  2. Andry Alamsyah, Akmal Wijaya Muharam, and Dian Puteri Ramadhani. 2024. Polarized narratives in digital spaces: A social network exami-Zicheng Cheng and Yanlin Li. 2024. Like, comment, and share on tiktok: Exploring the effect of sentiment and second-person view on the user https://jetstream-cloud.or… ↗
  3. Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. Roberta: A robustly optimized bert pre-training approach. find
  4. Tim Dettmers, Artidoro Pagnoni, Ari Holtzman, and Luke Zettlemoyer. 2023. Qlora: Efficient finetuning of quantized llms.
  5. Daniel Miehling, Daniel Dakota, and Sandra Kübler. 2025. Investigating polarization in youtube comments via aspect-based sentiment analysis. In Proceedings of the 15th International Conference on Recent Advances in Natural Language Processing - Natural Language Processing in the Generative AI era, p… find
  6. Timothy Dozat and Christopher Manning. 2017. Deep biaffine attention for neural dependency parsing. In 5th International Conference on Learning Representations (ICLR 2017), Toulon, France.
  7. Aoyu Gong, Sepehr Mousavi, Yiting Xia, and Savvas Zannettou. 2025. Clipmind: A framework for auditing short-format video recommendations using multimodal ai models. Proceedings of the International AAAI Conference on Web and Social Media, 19(1):671–687. find
  8. Jordan Miner and John E. Ortega. 2024. An NLP case study on predicting the before and after of the Ukraine–Russia and Hamas–Israel conflicts. In Proceedings of the Third Workshop on NLP for Positive Impact, pages 140–151, Miami, Florida, USA. find
  9. Benjamin Guinaudeau, Kevin Munger, and Fabio Votta. 2022. Fifteen seconds of fame: Tiktok and the supply side of social video. Computational Communication Research, 4(2):463–485. find
  10. Omkar Narkar, Aman Vohra, and Ashiqur R. KhudaBukhsh. 2025. Audience engagement with political messaging on youtube shorts (student abstract). Proceedings of the AAAI Conference on Artificial Intelligence, 39(28):29439–29441. find
  11. Pengcheng He, Jianfeng Gao, and Weizhu Chen. 2021. Debertav3: Improving deberta using electra-style pre-training with gradient-disentangled embedding sharing.
  12. Prabawati Nurhabibah, Hetilaniar Hetilaniar, Bernadus Wahyudi Joko Santoso, and Rustono Rustono. 2023. A multimodal critical discourse analysis of tiktok fyp video. International Journal of Social Science Research and Review, 6(2):208–216. find
  13. Gunther Jikeli. 2023. Holocaust distortions on social media after 10/7. the antisemitic mobilization. Technical report, Institute for the Study of Contemporary Antisemitism. find
  14. Alec Radford, Jong Wook Kim, Tao Xu, Greg Brockman, Christine Mcleavey, and Ilya Sutskever. 2023. Robust speech recognition via large-scale weak supervision. In Proceedings of the 40th International Conference on Machine Learning, volume 202 of Proceedings of Machine Learning Research, pages 28492–2…
  15. Alireza Karduni, Ryan Wesslen, Douglas Markant, and Wenwen Dou. 2023. Images, emotions, and credibility: Effect of emotional facial expressions on perceptions of news content bias and source credibility in social media. Proceedings of the International AAAI Conference on Web and Social Media, 17(1):… find
  16. Naquee Rizwan, Paramananda Bhaskar, Mithun Das, Swadhin Satyaprakash Majhi, Punyajoy Saha, and Animesh Mukherjee. 2025. Exploring the limits of zero shot vision language models for hate meme detection: The vulnerabilities and their interpretations. Proceedings of the International AAAI Conference on… find
  17. Rina Kumari, Nischal Ashok, Tirthankar Ghosal, and Asif Ekbal. 2021. Misinformation detection using multitask learning with mutual learning for novelty detection and emotion recognition. Information Processing & Management, 58(5):102631. find
  18. Naijia Liu, Xinlan Emily Hu, Yasemin Savas, Matthew A. Baum, Adam J. Berinsky, Allison J. B. Chaney, Christopher Lucas, Rei Mariman, Justin de Benedictis-Kessner, Andrew M. Guess, Dean Knox, and Brandon M. Stewart. 2025. Short-term exposure to filter-bubble recommendation systems has limited polariz… find
  19. Elisabeth Steffen. 2025. More than memes: A multimodal topic modeling approach to conspiracy theories on telegram. Proceedings of the International AAAI Conference on Web and Social Media, 19(1):1831–1844. find
  20. Phyo Yi Win Myint, Siaw Ling Lo, and Yuhao Zhang. 2024. Unveiling the dynamics of crisis events: Sentiment and emotion analysis via multi-task learning with attention mechanism and subject-based intent prediction. Information Processing & Management, 61(4):103695. find
  21. An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, Huan Lin, Jian Yang, Jianhong Tu, Jianwei Zhang, Jianxin Yang, Jiaxi Yang, Jingren Zhou, Junyang Lin, Kai Dang, Keming Lu, Keqin Bao, Kexin Yang, Le Yu, Mei Li, Mingfeng Xue, Pei… find
  22. Ellie Fan Yang, Lauren Kriss, and Yibing Sun. 2024. Fun with frustration? tiktok influencers’ emotional expression predicts user engagement with covid-19 vaccination messages. Health Communication, 39(12):2458––2473. find
  23. Heng Yang, Chen Zhang, and Ke Li. 2023. PyABSA: A modularized framework for reproducible aspect-based sentiment analysis. In Proceedings of the 32nd ACM International Conference on Information and Knowledge Management, CIKM ’23, page 5117–5122, New York, NY, USA. Association for Computing Machinery. find
  24. Wajdi Zaghouani, Mustafa Jarrar, Nizar Habash, Houda Bouamor, Imed Zitouni, Mona Diab, Samhaa El-Beltagy, and Muhammed AbuOdeh. 2024. The FIGNEWS shared task on news media narratives. In Proceedings of the Second Arabic Natural Language Processing Conference, pages 530–547, Bangkok, Thailand.
  25. Yang Zhang, Yawen Deng, Dong Wang, and Lanyu Shang. 2025. Multitec: A data-driven multimodal short video detection framework for healthcare misinformation on tiktok. IEEE Transactions on Big Data, 11(5):2471–2488. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/miehling-2026-multimodal-analysis-state-funded-news/