- Netflix is investing in technology to automate movie and television trailers, an executive said on a July earnings call.
- The technology could help Netflix lower the cost of creating trailers while adding personalization for its subscribers.
- In 2016, IBM’s Watson made a promo spot for the 2016 thriller “Morgan” by categorizing emotional moments in the film, although a human editor was still needed to polish it up.
(CBS) – The next movie trailer you watch on Netflix could be made by artificial intelligence.
Gregory Peters, chief product officer at Netflix, said in a July earnings call that the Los Gatos, California, company is investing in technology that can index characters and scenes in a movie “so that our trailer creators can really focus their time and energy on the creative process.”
It’s not the first time a tech company proposed taking artificial intelligence to Hollywood. In 2016, 20th Century Fox approached technology giant IBM to create the first “cognitive movie trailer” for its horror film “Morgan,” which is about a bioengineered human child who breaks free from her creators.
The machine learning system Watson assembled a trailer it considered “suspenseful” by analyzing the film visually to identify characters, objects and scenery, IBM scientists said at the time. It also consulted an internal database to identify emotions and interpret the mood of music to determine whether a scene was supposed to be eerie, frightening or tender.
The whole process took about 24 hours to complete after Watson “watched” the movie, indicating the growing aptitude of machines to categorize human emotion. Despite the progress, the “Morgan” trailer still required a human editor at the end of the process to polish a final version.
“Human beings are very quick to judge the mood of an individual,” Luke Scott, director of “Morgan,” said in a trailer spot explaining the technology. “Possibly an AI over time might be able to develop those same instincts.”
It remains difficult for artificial intelligence to create art, which requires the ability to originate thought. Trailer creators typically spend months editing cuts and commissioning original music to fine-tune the perfect promotion — and there’s a monetary reason behind all the effort.
“Trailers are commercials to a movie,” Wedbush analyst Michael Pachter said. “They are the last thing a real movie studio would cut corners on.”
Getting movie trailers to scale
But scientists as well as Hollywood executives say machine learning can help reduce the time required to make a trailer, as well as provide personalization.
AI “can certainly do 50% of the lifting,” said Matt Brubaker, CEO and creative director at Trailer Park, a Hollywood-based creative agency that specializes in movie and television trailers for clients, including Netflix. But he also asserted that “it won’t be as effective in the end as a trailer maker.”
According to IBM research scientist John R. Smith, IBM has already tapped its machine learning capabilities to help create highlight reels for major sports tournaments, like the Masters Tournament, U.S. Open and Wimbledon. Smith said that machines can parse through hours of footage to identify “the ingredients for an exciting moment,” like a crowd cheering or athletes lifting their arms for a high-five.
For Netflix, the ability to scale the creation and personalization of trailers would help improve its operating costs. It could also assist its in-house creative staff to develop personalized “thematics,” which are the auto video previews that play when scrolling through the streaming site.
“A lot of the trailer and highlight making process is one size fits all, everyone gets the same trailer,” Smith said. “But maybe you can get two or three trailers for different markets, and you can eventually make more personalized market.”
The technology company already pays top dollar to keep users engaged with its homepage, which Evercore analyst Lee Horowitz described as “probably one of the most valuable real estate online in terms of video content discovery.”
No “one-size” approach
Netflix personalizes the artwork shown to its 150 million subscribers when they tune in by relying on algorithms that can pick out which images may best resonate with them.
“Why should you care about any particular title we recommend? What can we say about a new and unfamiliar title that will pique your interest? How do we convince you that a title is worth watching?” Netflix researchers wrote in a 2017 Medium post.
The answer, they said, is to craft the perfect image to catch your eye on the homepage. “Then maybe, just maybe, you will give it a try,” the post continued.
A viewer watching romance movies may be more interested in watching “Good Will Hunting” if the artwork depicts a tender moment between actors Matt Damon and Minnie Driver, for instance. On the other hand, a comedy lover might be more likely to watch the film if Robin Williams is depicted.
All this personalization ultimately benefits Netflix’s bottom line. As Evercore analyst Horowitz noted, the longer a consumer remains on the platform, the more engaged they’ll be — and the more likely they are to move up to higher subscription tiers.