A Framework for Identifying Clickbait

The widespread use of clickbait on YouTube, fueled by sensational headlines, eye-catching thumbnails, and mislead- ing descriptions, has become a growing concern due to its deceptive nature and detrimental impact on user experience and content credibility. This manipulation of viewer attention undermines the trustworthiness of digital platforms and has prompted extensive research into automated detection techniques. In response, a variety of machine learning and deep learning approaches have been developed to identify clickbait effec- tively. These include natural language processing (NLP) methods for analyzing textual metadata, convolutional neural networks (CNNs) and vision transformers for detecting visual patterns in thumbnails, and sentiment analysis to evaluate the emotional tone of video titles and descriptions. Advanced transformer models such as BERT and GPT have shown significant success in capturing complex patterns across both text and image data, enhancing the accuracy of classification. Despite these technological advancements, evolving clickbait strategies continue to challenge detection systems. Future research should aim to improve the robustness, scalability, and real-time performance of these frameworks, with an emphasis on multimodal learning and cross-linguistic adaptability to better preserve the integrity and reliability of content-sharing platforms.