Netflix Aiming to Develop AI-Powered Movie Trailers

Consumers may soon be witness to movie trailers exclusively developed by an artificial intelligence engine. In their July earning’s report, Chief Product Officer of Netflix Gregory Peters informed the press that the company had invested money in a system that can be used to index characters and scenes from a movie. The system can then splice these scenes together to make a trailer that best represents the movie. While the system will have to go through typical learning time and iterations before it gets to the point of making human-like trailers, it’s a groundbreaking new way for the company to approach the trailer production process.

AI itself isn’t new to the trailer production process. In 2016, 20th Century Fox partnered with IBM to produce an AI-generated trailer for its horror flick “Morgan.” IBM’s Watson learned about the reactions humans had, cataloged the “scary” parts of the film and developed a trailer that still needed a human editor at the end to make it viewable. For a first attempt, it wasn’t bad, allowing the studio to have a rough draft trailer much faster than if a human trailer maker was developing the product.

Cutting Corners or Increasing Efficiency?

Movie trailers are the commercials that movie studios rely on to sell their products. Artificial intelligence still has problems developing artistic pursuits such as music and theatre. By introducing AI into trailer development, some analysts are concerned that it would result in a lower-quality trailer being produced in an attempt to cut corners. However, AI requires some time to build up a proper grasp of what it’s doing and learn from its mistakes. While the rough cuts of trailers today may need human editors, it’s very likely that machine learning will render those editors unnecessary within the next decade.

For Netflix, the draw of an AI would help improve their production costs, allowing them to spend less time and money on producing a workable trailer for a completed film. With the streaming service involved in creating a variety of shows for its audience, the advancement may significantly improve the company’s bottom line. Additionally, the company is developing personalized “thematics” and may eventually look into creating multiple trailers to appeal to different target demographics.

CloudWedge
Logo