Ben Affleck is an AI nerd, and the internet is impressed

Actor Ben Affleck drew attention this week after clips from recent interviews circulated, in which he discussed artificial intelligence with working knowledge of machine learning and film production. Appearing on GQ's "One More Question" series, he said he had been interested in computers since childhood and became more involved when analog film gave way to digital. He described writing what he called "pretty shitty Python scripts" and explained convolutional neural networks, the precursors to what a transformer can do, along with terms such as tensor, GPU and inference.

Affleck walked through how neural networks are used in visual effects workflows, including edge detection and feature extraction, which he described as identifying patterns well enough to know where a window ledge is so a green screen element can be replaced. He said he at times used his Hollywood profile to look at emerging AI technology, and that he visited OpenAI, which helped him see how to get more directly involved.

He addressed his startup's reported sale to Netflix earlier this year. Affleck said the $587 million figure is "not right" and noted he did not own the whole company. His stated method was to raise money, shoot for roughly eight months with many cameras and equipment, and build a proprietary dataset for late-stage training of open models on discrete tasks, with training code and inference code linked to specific work. Speaking separately with Bloomberg's Lucas Shaw at the Screentime 2026 conference in Los Angeles, he described taking open source models, unfreezing the weights, and training the last cinematic layer so a production crew could meet real production standards, while keeping proprietary work and gaining from the learning. He said he used the technique on his film "Animals," in post-production.

On AI risk, Affleck said his concerns center on his children in school, a 30% rise in A grades at colleges over three years, learned helplessness and responsible use. He said he does not worry about Skynet and doesn't think AI will take over the film business in any meaningful way, calling it additive.

Why it matters: The clips put a prominent filmmaker's account of open-model fine-tuning, custom datasets and production-grade AI pipelines into wide circulation, giving product and developer teams a public example of that workflow while a reported acquisition price remains disputed by the person involved.