AI may have been starring in sci-fi movies for decades, but time has come for it to leap from the screen into real life production rooms. Artificial intelligence (AI) in the film industry has come a long. Despite the ongoing battle of opinions, the majority of its usage is not "smart robots" and "terminators", but machine learning and cognitive services. And the possibilities are numerous.
In September 2016, 20th Century Fox revealed that it had used the AI capabilities of IBM's Watson to make a trailer for the sci-fi thriller Morgan. Not surprisingly, the buzz around this announcement was massive, with everyone wanting to see how well a machine had been able to complete such a creative and specialized task.
So, how did Watson's first shot at creating a Hollywood movie trailer turn out? Remarkably, it's a lot better than you’d probably expect. Aside from the opening narration line that comically refers to a person as "it," the trailer is surprisingly well done and serves as a fascinating indicator of the tremendous advancements that have been made in AI. At the same time, it's important to note that a real person edited the trailer, using the scenes that were selected by Watson, so AI didn't actually edit it. Regardless, Watson did learn about what makes horror trailers frightening by analyzing 100 samples, focusing on data points about visuals, audio and emotions to choose ten scenes from the movie to use in the trailer, and the results are pretty awe-inspiring. The event certainly marks a monumental moment in Hollywood history, illustrating one of the many ways in which artificial intelligence carries the potential to profoundly alter the film industry moving forward.
However, putting together a clip from the mass of existing material is not quite the same as creating something unique. So how far can machine learning go? Can machines imagine or create? It is a topic of equal interest and debate. Quite a few AI startups pitch the value of adding AI capabilities to creative arts. However, it's not quite time for script writers to begin looking for alternate employment, and the reality is that the creative process of making films will probably always be driven by human beings - or at least require human supervision.
At the same time, there are already many real-life applications for AI in the movie industry, such as automation to save time and increase efficiency in post-production, including automated syncing and grouping of clips. Additionally, machine learning is being heavily utilized for content analytics and personalized programming, adding relevant metadata, providing image analysis, and adapting to consumer preferences. Another area of high demand is predictive analytics and other tools for the marketing and advertising side of films. Startups like Vault, ScriptBook and Pilot Movies claim to predict how well any given movie will do at the box office by analyzing the script. Considering that creating the perfect formula for a blockbuster is a Hollywood dream, they may be here to stay.
Will Hollywood producers and film studios be using machines armed with AI to generate movie trailers in the near future? And even to write entire film scripts predicted to become instant hits? It seems somewhat far-fetched at this point, but there is no way to rule out these possibilities completely. There are undoubtedly many big players in the industry who are investigating this and other related prospects. One thing is certain - technology will be more and more involved in the process of creating movies on different levels, and artificial intelligence is yet to realize its full potential.