Artificial Intelligence in Media Entertainment and Sport 2025

Meta’s Segment Anything Model 2 (SAM 2) can significantly simplify labour-intensive content creation tasks by automating video and image editing. SAM 2 allows creators to quickly identify and isolate objects in both images and videos with its memory attention feature, which tracks objects across multiple frames. In post-production, SAM 2 can be used to make real-time adjustments like changing backgrounds or adding special effects without manually editing each frame.35CASE STUDY 9 Simplifying editing through intelligent image and video recognitionGenAI can significantly reduce costs and effort by streamlining production stages. “Co-pilots” can automatically generate clips, playlists and stories while enabling real-time metadata tagging, making it easier for creators to manage and enhance their content. For live events, genAI can help streamline advanced features such as instant slow-motion replays and dynamic camera configurations, accelerating the production of high-quality content at lower costs. The Deutscher Fußball-Bund (German Football Association) works with the sports management suite SAP Sports One’s AI co-pilot to enhance match preparation, automatically analysing data and extracting insights from opponents’ past matches. This reduces manual tasks for analysts and allows them to quickly respond to queries from team members preparing for a match, such as the opposing team’s game patterns, strengths and weaknesses.36 Accenture Song uses models trained on bespoke client taxonomies to generate and apply metadata to both new and existing assets at scale. This enhances asset use and value by improving their searchability, activation and analysis across the enterprise, potentially reducing related operating costs by up to 75%.37CASE STUDY 10 Powering sport performance adding co-pilots to the team CASE STUDY 11 Intelligent asset managementGenAI can use advanced models – such as image and object recognition – and natural language processing (NLP) to automatically enrich extracted metadata with semantic context, generating, for example, narrative patterns and scene descriptions. This supports human oversight activities, like content moderation, and reduces content processing time by automating repetitive tasks like metadata generation, tagging and content organization. Market researchers, for instance, can use genAI to optimize dataset processing and provide insights more accurately and efficiently.Content/asset creation Content/asset management Artificial Intelligence in Media, Entertainment and Sport 15