In the ever-evolving landscape of modern media, news agencies worldwide are spearheading the drive for unbiased and expansive multimedia news coverage across various languages. Emphasizing a commitment to both innovation and precision, the industry has seen the emergence of advanced AI-driven services that significantly enhance the accessibility and discoverability of video news content. This session explores how these tools are not just revolutionizing news access but are also exemplifying the ethical implementation of AI in journalism.
As AI becomes more embedded in news distribution, the necessity to identify and mitigate bias is paramount. Algorithms harboring biases can alter public narratives, deepen societal divisions, and inaccurately portray diverse populations. This presentation will outline the methods and strategies that leading news agencies are using to uphold impartiality in AI functionalities, addressing common issues such as accent, nationality, and gender biases in AI-driven content analysis and public figure recognition.
Further, the session will introduce the SHapley Additive exPlanations (SHAP) method, a robust technique for deciphering the decision-making processes within AI systems, particularly in the field of computer vision. This discussion is designed to merge the technical capabilities of AI with the ethical responsibilities associated with its use, providing participants with the necessary knowledge to promote transparency and accountability in the AI tools developed and utilized across the media sector.
Yulia Pavlova is the Head of Applied Innovation team at Reuters News, where she influences development and implements new AI-based technologies that transform the media and news organization through content creation, delivery, and distribution on a global scale.
Yulia Pavlova has 15+ years of experience in the R&D space and has 2 PhDs, one in Scientific Computing, and Applied Mathematics and Cybernetics.