Evaluation of Vision-Language Transformers for Multimodal News Authenticity and Integrity in Journalism: A Multi-Criteria Decision-Making Approach

In today’s digital age, inaccurate and invalid news spread may spread faster than fire. Evaluating textual and visual data is very important to stop this from happening. So, in Journalism, the Vision-Language Transformer (VLTs) are required to improve the verification of news integrity and authenticity throughout the Multimedia formats. The VLTs give us the unique benefit of recognizing false or altered information as they can process and examine data like texts and images at the same time. This approach involves accuracy, robustness, and situational symmetry to make the VLT model suitable for journalism. To assess VLT models’ suitability for journalism, this study looks at contextual alignment, accuracy, and robustness. The results highlight how VLTs may support trustworthy journalism by giving media outlets cutting-edge resources to uphold public confidence. This paper proposes a multi-attribute decision-making (MADM) framework and examines the VLT models on accuracy, flexibility, comprehensiveness, and reliability. The results in creating a potential VLT model to enhance the media integrity and authenticity.