Preparing for a keynote for the Barcelona VR360 conference recently, I used AI to get inspiration. I fed it the title of my talk “What will be the impact of AI on storytelling?”, and it responded:
The key to AI stories is not only that they will be able to be generated by an AI, but also be told in such a way that people will want to hear them. It will allow for an infinite amount of stories to be told in a multitude of ways, but only the best stories will survive. It is like the survival of the fittest in that it is survival of the best story.
The stories will be told in a variety of ways because the AI will be able to tell the story to the listener through a number of different mediums. If the AI is able to tell the story through a video game, it may choose to tell a story this way. If it can tell the story through a video, it may choose this medium. If it can tell it through a book, it may choose this medium. If it can tell it through a blog, it may choose this medium.
The AI will be able to choose the best medium, and the best way of telling the story, to tell the story. The AI will be able to use these different mediums to create a story that is not only a good story, but also the best story it can tell.
Not bad for an opener, and indeed pretty exciting as an idea for transmedia storytelling in its own right.
However, there is a big distinction between AI’s creation of a few conceptual paragraphs to the creation of a fully-fledged and engaging story.
Whether AI will achieve a full story is a goal which may or may not be attainable in the future, and indeed is a goal which is suspect in its intentions anyway.
In the shorter term, the current versions of AI have already enabled radically new opportunities for storytelling that break down restrictions imposed by the legacy of linear media - books, films, television and even, to some extent video games.
Current versions of AI - GPT-3 in particular - are triggering great innovation around four pillars of storytelling.
The first pillar is Time. Digital media in general removes the false time constraints of linear, analogue media. Linear media was an accident of technology. It was the best that the technology of the time could offer – books, film, television, video, audio. The constraints of a linear television show, the structure of a book, the length of a movie, all of these are limits on the technology that underpins them.
Most movies last around ninety minutes. A TV series might last ten hours. A game can be ten times that. With AI, we might have stories that last a year, ten years, and beyond. You could have a character who grows with you throughout your life. Now this is not new, perhaps because we are used to telenovelas and soap operas which have lasted throughout our lives. Expect many more long-growth storyarcs in the future, with characters that grow and age with us. Time is no longer a barrier to characters and stories.
Narratives are the second pillar. We are used to chapters, acts and scenes, but where the audience has influence, there needs to be a way of gauging that influence. Here we borrow from the games industry and “levels” to be able to chart the progress of the story. However my use here is an adaptation of the games industry’s language. It is not about making a story more difficult as it progresses, but more a way of identifying where the player is currently in the story. Perhaps we could also visualise a story as a sort of free-form board game of Snakes & Ladders with each section as a Square. Players can move through the Story Squares in a non-linear way, having different experiences each time. Indeed the way in which the player moves through the squares will precipitate a different story experience each time. For example, if the player interacts angrily with characters in Square Ten of the story, then when the player moves to Square Eleven with the same characters, naturally the story is going to be different than if they interacted warmly with those characters.
The third pillar is that new AI technologies, and specifically Machine Learning, create the potential for Infinite Stories. To clarify, these are not never-ending stories, but more than a story can have infinite pathways through it. Rather than everyone having identical experiences of s story, each version will be slightly different depending on our own individual input. The water cooler conversations are then matured from being: “Did you see that bit when…?” to “What I did at that bit when…?”. As well as being a more personal story, these infinite stories allow for replayability to see what would have happened if the audience - or a character - behaved differently at a certain moment.
Lastly, AI is a fabulous creative collaborator. It cannot write the full story, but platforms like GPT-3 provide rich, imaginative and surprising moments, thoughts and concepts which human creativity can then build upon. It cannot replace our own imagination which by definition is our own, as by definition a Rembrandt painting can never be created by AI, but it can speed up our inspiration, take us down pathways we would not have thought about, and reduce the fear of the blank page.
Here at Charisma, we learned a number of things in 2020 about how to write with AI which will be written up in a next post, but throughout all these developments, we know one key truth: Creativity should define Creative AI. The Creative Industries need to get excited about this new forms of storytelling to improve stories, reach new audiences, and maintain the pace of innovation that is running through television, theatre and games industries.
For the VR360 Conference, I read out the AI intro pretty much verbatim. But then moved on to my own narrative. It was a nice gimmick to introduce a talk and made the point well. And maybe we might use the ideas ourselves in the future.