Affleck details AI filmmaking workflow and model fine-tuning techniques

Ben Affleck, who sold his AI filmmaking startup to Netflix for a reported $587 million earlier this year, went viral for demonstrating technical knowledge of AI in interviews. In a GQ One More Question episode, he explained that visual effects workflows have included machine learning for years. He described using convolutional neural networks for edge detection and feature extraction, such as identifying window ledges to remove green screens. He noted he can write Python scripts and understands concepts like tensors, which are numerical translations of visual images including batch numbers, frame numbers, and RGB values. Affleck admitted using his Hollywood status to visit companies like OpenAI to understand emerging tech. He stated that to use AI ethically with artists, his startup created its own dataset by shooting for eight months with extensive equipment to serve as late-stage training for open models. At the Screen Time 2026 conference, he explained fine-tuning open-source models by unfreezing weights to meet cinematic standards, a technique used in his movie Animals for post-production. He clarified that the $587 million sale figure is inaccurate because he did not own the whole company.