Advertisement

Empathy machines: what will happen when robots learn to write film scripts?

A few years ago I moved to San Francisco, and almost everybody I met there immediately told me they were working on a startup. These startups all had the same innocent names – Swoon, Flow, Maker – and the same dreadful mission: to build AIs that automated some unfortunate human’s job. I always responded by pitching my own startup, Create. Create would build an AI that automated the creation of startups.

The tech bros never cared for my joke, but I did. In fact, I cared for it so much that I eventually began a novel about an android who wanted to become a screenwriter. It seemed an intriguingly comic premise, because unlike everybody else’s job, my job was clearly far too human to ever actually be automated.

So how would an android attempt to learn screenwriting? Generally, the best advice anybody can give an aspiring screenwriter is “don’t do it”. The next best advice is to watch a lot of movies.

But what would an android learn by watching the movies of 2050? (If you doubt we will have cinema-going androids in a few decades, count yourself lucky not to have spent much time amidst the tech bros.) Of course, by 2050 there may well only be a single movie released each year, a day-long Marvel-DC-Disney-Avatar crossover that will keep the merchandisers churning out toys year round. Our android could at least learn how to write a specific kind of plot from such a movie, as no doubt it would feature the same “mistreated-chosen-one-saves-the-world” storyline its spiritual forebears share today.

But mastery of plot might not get our android very far. Occasionally, I am sent scripts as potential rewrite jobs: they all have great plots, incredible talent attached, and a cover page that shows a half-dozen more in-demand screenwriters have already failed to fix them. Each time I sit down to read them, I am convinced I can be the chosen one that will save the day. But a few pages in, I invariably understand that no screenwriter can make this script work, because something crucial was missing at its conception: heart.

Roger Ebert famously described great movies as “machines that generate empathy”. I am hopelessly biased, but I suspect this empathy – this heart – can usually be traced back to the original storytelling animus. A movie that was conceived simply to make money, to win awards, or demonstrate somebody’s brilliance, will always lack the heart of a story that began with the primal human urge to share a story.

Assuming our android somehow had such a story he wished to share, how could he learn to write it from the heart? Well, perhaps by watching the same type of movie I like to believe taught me: the crowd-pleasing studio hits of the late 80s and early 90s. Certainly, these movies had their profound flaws: they were too white, too male and, nobody in them was ever worried about paying the rent. But something else undeniably unites them – unites Tim Robbins and Morgan Freeman on the beach in Mexico, Ethan Hawke climbing on his desk to address Robin Williams, and Kevin Costner standing in a cornfield communing with the ghosts of disgraced baseball players. And what it is that unites them is heart.

But how would an android even see those heartfelt old movies? They are already 30 years old, their best versions exist on physical film, and each year more of them fade, or combust, or are tossed into dumpsters as cinemas are converted to spin-bike studios. Their future looks bleak, and yet as a screenwriter I know that “all is lost” – the lowest point in a narrative arc, when the odds seem insurmountable – often precedes a triumphant finale.

So what if we deployed the well-worn screenwriting trick of reversal, and found a way to transform the very weakness of those beloved movies into their strength? What if on a particular day in, say, 2037, the entire world was for ever locked out of the internet? Might that not be a time when emotional movies stored on film regained their cultural supremacy over the endless onrushing gigabytes of chosen-one superheroes? Not by chance, it is in such a mis-topia – where humans have locked themselves out of the internet by forgetting the names of their favourite teacher and first pet – that my novel, Set My Heart to Five, takes place.

And one more screenwriting trick: the most satisfying endings to movies are often the ones that come full circle, yet still manage to surprise. Here, then, is our full circle and our surprise. Earlier this year, the Guardian published an article about the use of AI in movie-making entitled It’s a War Between Technology and a Donkey. As a dues-paid donkey, I could not bring myself to read it. I finally did so this morning, and turned pale at the line that said: “Within five years we’ll have scripts written by AI that you would think are better than human writing.”

So, surprise: my job is being automated after all. Within five years – no doubt courtesy of a startup called Scripter or Draft – screenwriting AIs will be here. It takes an average of seven years to get a movie made, so what is the point in even beginning a new screenplay now? So, no more screenwriting for me. But if you’d like to hear about exciting opportunities to become an early investor in an exciting startup called Create, please do get in touch.

• Set My Heart to Five by Simon Stephenson is published by 4th Estate.