It should come as no surprise that an artificial intelligence application is writing horror prose. AI bots have spread their wings in recent years, challenging the premise that automated AI technology will be confined to chess, analytics, or number-crunching. Indeed, they have been recording music, writing movies, and dabbling in the visual arts. This all squares perfectly with author Ray Kurzweil’s prediction that AI — and transhumanist AI augmentations — would facilitate exponential improvements in every imaginable human endeavor, including making better art.
Now, just in time for this year’s Halloween comes news that a deep learning algorithm named Shelley AI (named after Frankenstein author Mary Shelley) is producing short horror stories — and they’re creepy as hell. Shelly AI uses ideas submitted via Twitter and a subreddit devoted to short horror and spins prose off these snippets of text. According to researchers at MIT, who developed the algorithm, Shelley AI’s writing “takes a bit of inspiration in the form of a random seed, or a short snippet of text, and starts creating stories emanating from her creepy creative mind.”
One passage reads:
I slowly moved my head away from the shower curtain and saw the reflection of the face of a tall man who looked like he was looking in the mirror in my room. I still couldn’t see his face, but I could just see his reflection in the mirror. He moved toward me in the mirror, and he was taller than I had ever seen. His skin was pale, and he had a long beard.
Shelley continued the story using a Twitter submission.
“Please, don’t be scared, I’m just looking for her,” he said pointing to the mirror. I nodded, frozen, because something touched me. I remember he looked down at me and screamed in agony. He didn’t say anything, but he looked over at me with that unnatural expression.
As the Shelley AI website describes, the operating principle seems to be human-AI collaboration. Ahead of last year’s Halloween, the team behind the algorithm developed Nightmare Machine, AI-generated horror imagery designed to both analyze and elicit extreme emotions in human observers. This year they asked an almost hauntingly prophetic question related to artificial intelligence: “Can machines learn to scare us?”
You tell us. Other fragments of Shelly AI’s writing include the following:
I could feel something watching me and I could see this dark shadow standing there with a torrent of hatred in its face. I was beyond scared so I didn’t take my eyes off this thing and turned back towards the mirror to see what was behind me!
Another excerpt of a story, which you can find more of here, continues below:
She fell to the floor from her cries and muttered a soft “Come to meeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee”. He stood rooted with fear, she slowly crawled spider-like towards him, and started shrieking ‘no escape nowhere to hide’.
There seems to be an almost perfectly calibrated creepiness endowed to these stories. Years ago, shortly before any deep learning algorithms were producing actual art, I wrote a fake review of the first artificial intelligence-directed horror movie. In the review, I positioned the movie as a blockbuster flop, and the director, Art 5, as a studio hack that was using every cliche in the book to manufacture a tired zombie apocalypse film. Maybe I wasn’t giving enough credit to how deeply embedded into our subconscious AI would become, nor how knowledgeable machines would be of our fears.