Increasing business results through AI in the entertainment industry
summary
Crowdsourced training on 300 episodes of Friends labels about 80,000 faces tagged by talent, emotion and intensity, then trains a pipeline that uses one talent model + five neural nets for emotion detection. The system scans catalogs to extract high-emotion moments, maps those moments to core audience segments, and runs experiments measuring impact on intent to view and purchase. Outcome: automated, scalable generation of targeted marketing creative from actual video moments.
AI-written summary · about
related talks
20mThe Future Is Agentic: Unlocking Creative Research With AI
37mAI and Emotions: opportunities and challenges (Virtual Talk)
22mHarnessing the Potential: AI's Role in Enchanting Media Experiences
21mSession: AI x UA – Redefining Creative Strategy For The Next Decade
16mSession: AI Video Ads – Built 90% Faster & 80% Cheaper
19mAI video is Already App-ening
18mCommercial applications of AI
39mVideo Generation + AI: From Production to Product