Designing a Multimodal Hate Speech Detection Model For X Platform: A Systematic Analysis of Current Approaches

The proliferation of hate speech on social media platforms poses significant societal challenges, with X platform experiencing a 50% overall increase in hate speech, including a 260% rise in transphobic slurs following recent policy changes. Traditional text-based detection models struggle with modern communication patterns, particularly on platforms like X where the 280-character constraint encourages coded language and linguistic compression. This study addresses critical gaps in multimodal hate speech detection through two primary objectives: systematic analysis of current multimodal models to identify gaps and limitations specific to X platform, and to design of an innovative multimodal architecture optimized for X platform's unique communication environment. This analysis of six prominent models—VisualBERT, UNITER, HGAT, Stacked Ensemble Framework, Multimodal Transformers, and Visual Data Augmentation approaches—reveals that zero percent of existing models address X platform's 280-character patterns, 83% show text over-reliance, and all models fail real-time processing requirements (<500ms). These findings provide performance analysis and gap analysis across multiple evaluation dimensions. In response, this study presents a designed a novel six-layer architecture featuring breakthrough Dynamic Cross-Modal Attention mechanisms, compression-aware text processing, and lightweight vision transformers specifically optimized for X platform. The architectural design addresses identified gaps through platform-specific preprocessing, parallel feature encoding across four specialized components (Platform-Optimized RoBERTa, Lightweight Vision Transformer, Cultural Context Analyzer, and Adaptive Learning Module), and dynamic multimodal fusion achieving balanced processing between textual and visual modalities. This research contributes to advancing hate speech detection methodologies by providing gap analysis and presenting an innovative design framework that addresses real-time processing, platform-specific optimization, and balanced multimodal integration which are critical requirements for practical social media content moderation.

Multimodal hate speech detection, X platform

1. Adewumi, T., Liwicki, M., Alfter, D., Abera, N. S., & Alabi, J. (2024). Fairness analysis in multimodal hate speech detection models. Journal of AI Ethics, 15(2), 145–162. [Google Scholar] [Crossref]

2. Agarwal, S., & Chowdary, A. (2021). Deep learning approaches for hate speech detection: A comprehensive survey. ACM Computing Surveys, 54(8), Article 163. https://doi.org/10.1145/3457607 [Google Scholar] [Crossref]

3. Ahmed, F., & Lee, K. (2024). UNITER-based multimodal hate speech detection on social media. IEEE Transactions on Computational Social Systems, 11(3), 234–248. [Google Scholar] [Crossref]

4. Alemayehu, M., Yimam, S. M., & Biemann, C. (2024). Multilingual hate speech detection challenges in Ethiopia. African Journal of Information and Communication, 29, 45–63. [Google Scholar] [Crossref]

5. Alsaedi, N., Burnap, P., & Rana, O. (2023). BERT-based models for hate speech detection: Performance analysis. Natural Language Engineering, 29(4), 891–910. [Google Scholar] [Crossref]

6. Andrew, S. (2024). X platform transparency report 2024: Hate speech statistics. Platform Safety Reports, 12, 78–95. [Google Scholar] [Crossref]

7. Arya, V., Sethi, A., & Verma, A. (2024). CLIP for multimodal hate speech detection in memes. Computer Vision and Image Understanding, 238, Article 103847. [Google Scholar] [Crossref]

8. Ayetiran, E., & Özgöbek, Ö. (2024). Inter-modal attention mechanisms for hate speech detection. Pattern Recognition Letters, 178, 112–119. [Google Scholar] [Crossref]

9. Bhattacharya, S., Singh, S., Kumar, R., Bansal, A., Bhagat, A., Dawer, Y., Lahiri, B., & Medya, S. (2023). Graph neural networks for detecting hate speech clusters in online communities. Social Network Analysis and Mining, 13, Article 89. [Google Scholar] [Crossref]

10. Bose, A., Hamilton, W., & Guha, N. (2023). Cross-attention layers in VisualBERT for multimodal hate speech detection. IEEE Access, 11, 45231–45244. [Google Scholar] [Crossref]

11. Center for Countering Digital Hate. (2023). Failure to protect: Twitter’s hate speech problem. https://www.counterhate.com/ [Google Scholar] [Crossref]

12. Charles, M. (2024). X platform user statistics and trends 2023–2024. Digital Media Analytics, 18(2), 156–171. [Google Scholar] [Crossref]

13. Chakma, K. (2024). Semantic role labeling for enhanced hate speech detection. Computational Linguistics, 50(1), 89–112. [Google Scholar] [Crossref]

14. Chen, Y. C., Li, L., Yu, L., El Kholy, A., Ahmed, F., Gan, Z., Cheng, Y., & Liu, J. (2020). UNITER: Universal image-text representation learning. In A. Vedaldi, H. Bischof, T. Brox, & J. M. Frahm (Eds.), Computer Vision – ECCV 2020 (pp. 104–120). Springer. [Google Scholar] [Crossref]

15. Cohen, D., Freedman, M., & Shahaf, D. (2024). VisualBERT applications in hate speech detection. Journal of Machine Learning Research, 25, 1847–1872. [Google Scholar] [Crossref]

16. Corso, G., Cavalleri, L., Beaini, D., Liò, P., & Veličković, P. (2024). Heterogeneous graph attention networks for multimodal hate speech detection. Proceedings of the AAAI Conference on Artificial Intelligence, 38(11), 11234–11242. [Google Scholar] [Crossref]

17. Cui, L., Lee, D., & Huang, B. (2023). Majority voting ensemble for hate speech detection. ACM Transactions on Intelligent Systems and Technology, 14(3), Article 42. [Google Scholar] [Crossref]

18. Devi, S., Singh, R., & Sharma, A. (2024). Hierarchical attention networks for text-based hate speech detection. Information Processing & Management, 61(2), Article 103245. [Google Scholar] [Crossref]

19. Dey, S., Chakraborty, T., & Naskar, S. K. (2024). XGBoost for nuanced hate speech detection. Expert Systems with Applications, 237, Article 121456. [Google Scholar] [Crossref]

20. Duong, V., Liang, P., Nguyen, T., & Tesillo, R. (2022). Heterogeneous graph attention networks for hate speech detection on social networks. IEEE Transactions on Knowledge and Data Engineering, 34(8), 3845–3858. [Google Scholar] [Crossref]

21. Gachara, M., & Gachara, W. (2024). Hate speech challenges in Kenya’s digital landscape. East African Journal of Information Technology, 6(1), 23–41. [Google Scholar] [Crossref]

22. García-Hidalgo, I., Aparicio, F., Moya-Alcover, G., & Buades, A. (2024). Data augmentation techniques for visual hate speech detection. Image and Vision Computing, 142, Article 104892. [Google Scholar] [Crossref]

23. Gomez, R., Gibert, J., Gomez, L., & Karatzas, D. (2020). Exploring hate speech detection in multimodal publications. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (pp. 1470–1478). IEEE. [Google Scholar] [Crossref]

24. Gong, H., Mu, Y., Li, Q., & Zhang, X. (2024). Late fusion strategies for multimodal hate speech detection. IEEE Transactions on Multimedia, 26, 3456–3468. [Google Scholar] [Crossref]

25. Harry, C., & Heng, S. (2024). Adversarial robustness in multimodal hate speech detection. International Journal of Computer Vision, 132(4), 1234–1251. [Google Scholar] [Crossref]

26. Hashim, I., Alsmadi, I., & Al-Ayyoub, M. (2024). Multimodal transformers for contextual hate speech analysis. Pattern Recognition, 147, Article 109876. [Google Scholar] [Crossref]

27. Hebert, D., Martinez, A., & Chen, L. (2024). Advanced multimodal transformer architectures for hate speech detection. Neural Computing and Applications, 36(8), 4123–4142. [Google Scholar] [Crossref]

28. Huang, F., Zhang, X., & Li, Z. (2022). Multi-head attention in ensemble frameworks for hate speech detection. IEEE Transactions on Neural Networks and Learning Systems, 33(12), 7234–7248. [Google Scholar] [Crossref]

29. Ivan, A., Martinc, M., Pelicon, A., Purver, M., & Pollak, S. (2024). Multimodal fusion techniques for social media content analysis. ACM Computing Surveys, 56(5), Article 123. [Google Scholar] [Crossref]

30. Jahan, M. S., & Oussalah, M. (2023). A systematic review of hate speech automatic detection using natural language processing. Neurocomputing, 546, Article 126284. [Google Scholar] [Crossref]

31. Jiang, T., Li, J., Hovy, E., Chang, K. W., & Peng, N. (2024). BERT optimization for hate speech detection tasks. IEEE Access, 12, 23451–23466. [Google Scholar] [Crossref]

32. Julia, C. (2025). Analysis of hate speech trends on X platform post-acquisition. Digital Society Research, 7(1), 34–52. [Google Scholar] [Crossref]

33. Kara, S. (2025). The impact of policy changes on hate speech prevalence on X. Journal of Online Safety, 13(2), 67–85. [Google Scholar] [Crossref]

34. Ketineni, S., & Jayachandran, R. (2024). Dynamic attention mechanisms for multimodal hate speech detection. Expert Systems with Applications, 241, Article 122645. [Google Scholar] [Crossref]

35. Kim, S., Park, J., & Lee, H. (2025). Visual data augmentation strategies for robust hate speech detection. Computer Vision and Image Understanding, 240, Article 103912. [Google Scholar] [Crossref]

36. Lakzaei, M., Ramezani, M., & Rahmani, H. (2024). Graph convolutional networks for social media hate speech detection. IEEE Transactions on Computational Social Systems, 11(2), 891–904. [Google Scholar] [Crossref]

37. Li, L. H., Yatskar, M., Yin, D., Hsieh, C. J., & Chang, K. W. (2019). VisualBERT: A simple and performant baseline for vision and language. arXiv preprint. https://arxiv.org/abs/1908.03557 [Google Scholar] [Crossref]

38. Li, X., Wang, Y., & Chen, M. (2024). Multimodal alignment score for hate speech detection evaluation. ACM Transactions on Multimedia Computing, Communications and Applications, 20(4), Article 112. [Google Scholar] [Crossref]

39. Lu, Y., Zhang, C., & Tang, R. (2025). Advanced OCR techniques for social media text extraction. International Journal of Document Analysis and Recognition, 28(1), 45–62. [Google Scholar] [Crossref]

40. Ma, J., Gao, W., & Wong, K. F. (2022). Cross-dataset evaluation of multimodal hate speech detection models. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (pp. 3456–3468). Association for Computational Linguistics. [Google Scholar] [Crossref]

41. Mahajan, R., Gupta, D., & Sharma, P. (2024). Stacked ensemble frameworks for multimodal hate speech detection. Pattern Recognition Letters, 177, 89–97. [Google Scholar] [Crossref]

42. Mao, R., Liu, Q., He, K., Li, W., & Cambria, E. (2025). Cross-modal consistency in multimodal hate speech detection. IEEE Transactions on Affective Computing, 16(1), 123–137. [Google Scholar] [Crossref]

43. Mim, S. (2024). Meta-classifier approaches in stacked ensemble hate speech detection. Machine Learning with Applications, 15, Article 100487. [Google Scholar] [Crossref]

44. Mody, S., Kharosekar, A., Puranik, K., Aroskar, T., Srivastava, A., Vyawahare, H., Kiwelekar, A. W., & Netak, L. D. (2023). Hate speech detection: Challenges and opportunities. ACM Computing Surveys, 55(13s), Article 269. [Google Scholar] [Crossref]

45. Patel, D., Singh, P., & Thakkar, A. (2023). ViLBERT for multimodal hate speech detection: Capabilities and limitations. Computer Vision and Image Understanding, 228, Article 103612. [Google Scholar] [Crossref]

46. Salman, A., Khan, W., & Ali, S. (2024). Random forest ensemble for balanced hate speech detection. Applied Soft Computing, 152, Article 111234. [Google Scholar] [Crossref]

47. Saunders, M., Lewis, P., & Thornhill, A. (2019). Research methods for business students (8th ed.). Pearson Education. [Google Scholar] [Crossref]

48. Santos, T., Oliveira, H. P., & Cunha, A. (2024). SHAP values for multimodal feature importance analysis in hate speech detection. Explainable AI, 8(2), 156–174. [Google Scholar] [Crossref]

49. Siddiqui, S., Kumar, A., & Jain, A. (2024). Self-attention mechanisms for implicit hate speech detection. Natural Language Engineering, 30(2), 345–368. [Google Scholar] [Crossref]

50. Soares, F., Villavicencio, A., & Pardo, T. A. S. (2024). Coded language and euphemisms in compressed social media text. Journal of Computational Linguistics, 48(3), 567–589. [Google Scholar] [Crossref]

51. Statista. (2024). Meta’s hate speech content removal statistics Q1 2024. https://www.statista.com/ [Google Scholar] [Crossref]

52. Tsiamas, I., Papadopoulos, S., & Kompatsiaris, Y. (2025). Compression-aware natural language processing for microblogging platforms. Computational Linguistics, 51(1), 78–102. [Google Scholar] [Crossref]

53. Tyagi, A., & Szénási, S. (2023). Cross-modal attention networks for hate speech detection. Neural Processing Letters, 55, 3421–3438. [Google Scholar] [Crossref]

54. Velioglu, R., & Rose, J. (2020). Detecting hate speech in multimodal memes. arXiv preprint. https://arxiv.org/abs/2012.14891 [Google Scholar] [Crossref]

55. Xue, H., Chen, Y., & Li, M. (2025). Temporal synchronization in multimodal content processing. IEEE Transactions on Multimedia, 27(2), 234–249. [Google Scholar] [Crossref]

56. Yadav, A., Vishwakarma, D. K., & Kumar, A. (2021). Limitations of traditional machine learning in hate speech detection. Expert Systems, 38(6), Article e12701. [Google Scholar] [Crossref]

57. Yang, L., Zhang, H., & Wang, X. (2025). Semantic correlation analysis in multimodal hate speech. Pattern Recognition, 149, Article 110234. [Google Scholar] [Crossref]

58. Zeng, W. (2024). Visual data augmentation for improved hate speech detection. IEEE Access, 12, 89234–89248. [Google Scholar] [Crossref]

59. Zhou, L., Palangi, H., Zhang, L., Hu, H., Corso, J., & Gao, J. (2020). Unified vision-language pre-training for image captioning and VQA. Proceedings of the AAAI Conference on Artificial Intelligence, 34(7), 13041–13049. [Google Scholar] [Crossref]

• The Role of Artificial Intelligence in Revolutionizing Library Services in Nairobi: Ethical Implications and Future Trends in User Interaction
• ESPYREAL: A Mobile Based Multi-Currency Identifier for Visually Impaired Individuals Using Convolutional Neural Network
• Comparative Analysis of AI-Driven IoT-Based Smart Agriculture Platforms with Blockchain-Enabled Marketplaces
• AI-Based Dish Recommender System for Reducing Fruit Waste through Spoilage Detection and Ripeness Assessment
• SEA-TALK: An AI-Powered Voice Translator and Southeast Asian Dialects Recognition