Sunspring’s first debut was in Sci-fi London Film Festival 2016; the story was set in a miserable world with extremely high unemployment rate, which attracted a certain number of fans. However, the most surprising feature about the film was its screenwriter – an Artificial Intelligence bot.
One of the common reviews about the movie was “amusing but strange”, it was spotted in character’s dialogues that the sentences rather sound unrelated and random. Nevertheless, AI’s work on background music and visual images were excellent. This spurred the idea of human- machine collaboration in storytelling.
The Massachusetts Institute of Technology (MIT) Media Lab recently looked into the potential for this combination in video storytelling. Before getting into the research, it is found that a story’s emotional arc affects the audience engagement greatly. The most fascinating story usually provokes strong emotional response.
By working with McKinsey’s Consumer Tech and Media team to program a machine-learning algorithm to be a moviegoer. Using deep neural networks, it estimates positive or negative emotional content by second harvested from short videos, movies, TV… “Watching” the opening of Up – an animated film which was a popular hit, the computer drew a graph of emotional arc based on its visual effects and sound.
After having the model watched thousands of videos and constructing emotional arcs for each one, researchers classified the stories into families of arcs. Among more than 500 Hollywood movies and another with almost 1,500 short films found on Vimeo, the data was divided into 5 groups of families. The next step was to analyze and predict the audience engagement on social media. After running on test on several of families, two stories that attracted the most engagement both culminate with a positive emotional bang found in the climax. Hence, it could be concluded that positive emotions generate greater engagement. Not only are there more in number of engagement but these comments are also longer and passionate as well.
In general, the idea of human collaborating with machine is no longer science fiction. But it also requires regulations for this co-operation. Developing rules and ethical guidelines for AI are what the AIWS is working on, with the development of Layer 1, AIWS hopes to establish a responsible code of conduct for AI Citizens to ensure that AI is safely integrated into human society.