Image courtesy 20th Century Fox

Lights! Camera! Action! IBM Watson Directs Morgan, its first Movie Trailer

Cognitive computing decomposes movie and suggests scenes

IBM Watson has done many cool things; from competing with humans on a game show to whipping up bespoke recipes. And now Watson has its movie-making debut advising the Director on the trailer for the 20th Century Fox AI (artificial intelligence) thriller/horror film Morgan.

Fox wanted to explore whether it can use AI to create a horror movie trailer for its upcoming film about a terrifying AI character named Morgan. Morgan ends up learning and developing too quickly for her own good and goes all Frankenstein on her creators.

The challenge is to not only teach the Watson to understand, “what is scary,” but then to create a trailer that is considered “frightening and suspenseful” by viewers.

IBM Research fed Watson more than 100 horror film trailers cut into separate moments and scenes. After learning what keeps audiences on the edge of their seats, Watson used tone and emotion analytics to understand each scene in the entire movie. It then suggested ten scenes from Morgan for the editor to include in the trailer.

Using the power of cognitive computing cut the creation process down to only 24 hours when it typically takes around 10 to 30 days to complete a trailer.

Beyond the time-saving this project also shows the depth of analytics now possible with cognitive technologies. We can now identify certain qualities in media, like their ability to elicit certain emotional responses, such as anger, disgust, fear and sadness. Outside of the film industry, given that so much of business is driven by communications, the potential here is significant. For instance, think of how a call center can use this same form of cognitive analysis applied to customer calls to better understand how a new product is perceived by the market. Are the people that call up generally angry or frustrated?

You can read what Wired and Business Insider had to say about the Morgan trailer, and this more in-depth IBM Research blog post.