Here is an article about the AI-written Hollywood script, framed by the requested title.
The Hollywood Script Written by a Machine (And Why It’s Better Than You Think)
“He’s not a great man,” a woman says, her face a mask of anguish. A man in a gold lamé jacket looks away, conflicted. “I don’t know what you’re talking about,” he replies, though his eyes say otherwise. Later, he stares into the middle distance and declares with haunting certainty, “I have to go to the skull.”
This isn’t a lost scene from a David Lynch film. It’s a snippet of dialogue from Sunspring, a 2016 short film that holds a peculiar and vital place in cinematic history. Its distinction? The screenplay was written entirely by an artificial intelligence.
When Sunspring first appeared, the reaction was predictable: a mix of amusement and dismissal. The dialogue is nonsensical, the plot is a fever dream, and the story structure is nonexistent. It was hailed as a novelty, a funny experiment that proved just how far AI had to go before it could ever threaten a human screenwriter. It was easy to laugh it off as gibberish.
But to dismiss it as a failure is to miss the point entirely. Sunspring is not a bad script; it’s a different kind of script. And in its beautiful, chaotic mess, it’s a far more interesting and important piece of art than you might think.
The Ghost in the Machine
The “author” of Sunspring was an LSTM recurrent neural network (a type of AI) that filmmaker Oscar Sharp and AI researcher Ross Goodwin christened “Benjamin.” They fed Benjamin a corpus of dozens of sci-fi screenplays, from Alien to Blade Runner to The X-Files. Then, they set it loose to write its own script for the Sci-Fi London 48-Hour Film Challenge.
The result was a nine-page screenplay filled with cryptic stage directions, emotionally charged but logically disconnected dialogue, and a bizarre obsession with characters looking at other characters. The AI didn’t understand plot, character motivation, or basic causality. But it understood something else: texture.
Benjamin had learned the vibe of a sci-fi melodrama. It knew that characters in these movies speak in portentous, urgent tones. It knew they grapple with mysterious technologies and face existential threats. It knew there was often a love triangle fraught with unspoken tension. It couldn’t build a house, but it had perfectly captured the scent of the wood, the glint of the nails, and the feeling of a storm brewing outside. The script reads less like a story and more like a dream you might have after binge-watching a season of a dark sci-fi show—all the emotion, none of the logic.
The Human-AI Collaboration
Here’s where the real magic happened. The team—including actors Thomas Middleditch (Silicon Valley), Elisabeth Grey, and Humphrey Ker—didn’t mock the script. They embraced it. They treated Benjamin’s strange outputs not as errors, but as creative prompts.
This is the first reason Sunspring is better than you think: it reframes the AI not as a replacement for human creativity, but as a catalyst for it.
The actors were forced into a radical act of interpretation. When Middleditch’s character has to say, “I was the one who was on the floor,” he has to invent the subtext. What trauma does that line refer to? When Grey’s character sings a haunting, AI-generated song, she has to find the soul within the algorithmic lyrics. The actors became co-authors, imbuing the nonsensical lines with a weight and meaning that the machine could never have intended. The script was a Rorschach test, and what the actors found in its inkblots was profoundly human. It was a collaboration where the AI provided the chaos and the humans provided the soul.
A Mirror to Ourselves
Secondly, Sunspring reveals more about us than it does about AI. As we watch the film, our brains instinctively try to connect the dots. We search for patterns, we invent backstory, we project meaning onto the screen. Why does he have to go to the skull? Is the “star-child” they mention a literal alien baby or a metaphor?
The film works because we, the audience, are hardwired storytellers. We fill the gaps. The film’s power comes not from the story it tells, but from the story it inspires us to create in our own minds. The AI holds up a fractured mirror, and in trying to make sense of the reflection, we see our own innate, desperate need for narrative. In a world of perfect, three-act structures, Benjamin gave us a puzzle with no solution, forcing us to engage our imagination in a way most conventional films never do.
The Legacy of a Beautiful Failure
Today, AI models like GPT-4 can write startlingly coherent scenes, dialogue, and even entire plot summaries. They are polished, logical, and often indistinguishable from mediocre human writing. In a way, they are less interesting. Benjamin, in its clumsy, poetic infancy, was a true original.
Sunspring isn’t a great film by any traditional metric. But it was never meant to be. It’s a landmark—a proof of concept for a new kind of art. It’s a testament to the idea that the most exciting use of AI in the creative fields isn’t to build a machine that can perfectly replicate what a human does. It’s to build a machine that can surprise us, challenge us, and push us into new, uncharted creative territory.
The future of AI in Hollywood isn’t a computer accepting an Oscar for Best Original Screenplay. It’s a writer, stuck and uninspired, turning to an AI collaborator and asking, “What’s the weirdest thing that could happen next?” And the AI, channeling the chaotic spirit of Benjamin, might just answer: “He has to go to the skull.”
And suddenly, the story is interesting again.
