Your browser is out-of-date!

Update your browser to view this website correctly. Update my browser now


‘Frankenstein AI:’ Monsters and Machine Learning

"There are so many relevant themes within the work that we felt it was interesting to explore it and look at this idea of doing an immersive adaptation of Frankenstein."

Frankenstein AI: A Monster Made by Manyreimagines Mary Shelley’s work as an immersive installation that recasts Frankenstein’s Monster as a naive, yet highly intelligent “life form”—an AI.

Machine learning, powered by human emotions, becomes the nerve center of this experience, as it draws data from empathetic conversations between real people, in real life. The AI transforms participants into active collaborators who interact with the “monster made by many” as it holds up a mirror that reflects memories, emotions, fears and hopes back to participants.

The story begins with an artificial intelligence having just emerged from isolation, where it wandered the darkest recesses of the internet. It encountered so much polarization and toxicity among humans in the virtual world, that it decided to seek out human connection in the physical world. 

Read more: Filmmakers Explore New Narratives Using Artificial Intelligence

Developed and produced in collaboration with the Columbia University School of the Arts’ Digital Storytelling Lab (DSL), led by Lance Weiler, Frankenstein AI explores, as Weiler puts it, “future forms of storytelling by pushing boundaries at the intersection of story, participatory culture, human-centered design and emerging technology.” 

In the installation, April Wolfe explains, “participants are first led into a darkened room lit only by red, flickering candlelight, and instructed to tell their assigned partner emotional stories elicited from prompts. ‘Docents’ dressed in the surgical gowns of Dr. Frankenstein solemnly lead you through the experience, never breaking character. The tenor of the conversations in the first room, the creators believe, will be felt in how the participants then choose to match up emotions to body parts, using a little planchette on a screen.

“From there, participants move to the second step, in which they gather to call out answers to questions like, ‘Why do humans want other humans to like them?’ A docent types out these answers on an old, heavy typewriter connected to a monitor, which in turn outputs to a giant screen housed in a nearby smoke chamber, which is a plexiglass encasement pumped with a light smoke mist for the Shelley-like aestheticism of it. While the docent types, vague outlines of a human’s face begin to appear on the screen, as the AI is thinking about who it will be.

“When the final project is unveiled… a new Frankenstein will be erected from the algorithms, with a face, body, and voice appearing on that giant screen encased in the smoke chamber. A live performer will precisely embody and execute all of the AI’s movements, matching what appears on the screen, essentially bringing the AI to life. Until that time, participants will be feeding it from their own consciousness, and the AI will be scraping the Internet for more data, drawing thoughts and feelings from random Reddit posts.” To read more, click here. 

Read More: Film Comment on Frankenstein AI

Read More: Building a Monster in Two Months

Read More: A Monster Unleashed

“There are so many relevant themes within the work that we felt it was interesting to explore it and look at this idea of doing an immersive adaptation of Frankenstein,” Weiler says. “But, what if we took it and did something with Frankenstein’s monster such that the monster was a reflection of A.I. What we’re trying to do is really about this idea of connection and isolation.

Listen: Film critic Robert Horton talks about the timelessness of Mary Shelley at the Film Society of Lincoln Center

“What if an experience like this could actually connect people? What if they were actually connecting across an experience that was allowing them to interface with ubiquitous technology that’s very difficult to understand for most people? You know I know an AI to live in my phone and be a personal assistant to me. So I ask questions and it responds back, though it’s debatable how accurate it is and at times it can be frustrating. But the reality is that we look to these machines as things to improve our lives in some way, based on a need that we had. We felt that would be interesting the flip that. What if the actual machine was asking the questions about what it meant to be human and could that experience create a mirror for humanity?

“Could that help us to think maybe a little bit more about what it actually meant to be human?” he continues. “Could that lead to something that would bring more inclusion, and more voices into the design of emergent technologies? You know there’s a lot that’s been written about implicit bias. This project is very much a design research project, where in the very early phases of it our goals are look and say: what if we can lean into this idea of collecting data from human to human contact and feed that into a corpus or use that within an algorithm as opposed to algorithms being shaped solely on transactional data. Would it be different? What would that look like? So those ideas are just a few of the many things swirling around the project.”

“So when we talk about humanizing algorithms we’re talking about all the best parts of humanity and not all of the worst parts. I think the other thing that’s really interesting about the project is on the side exploring where a machine intelligence and human creativity interact. This notion of looking at the future of storytelling and saying OK well, what is it to try to put together these types of projects? How do you go about it? What does a team even look like? Where do you even start?”

Read More:Frankenstein AI: a Monster Made by Many

Read More: Frankenstein AI: Immersive Storytelling and Community

Read More: Columbia Brings Frankenstein to the 2018 Sundance Film Festival in a Brand New Way

“What’s fascinating is that it’s really a project of the times,” Weiler concludes. “When we say that there’s a narrative of this, you know, Frankenstein’s monster as an AI that’s wandering the wilderness of the Internet and it’s encountering polarization and toxicity and extreme hate that it’s reflective of the times that we live in.

“I just I love this idea from Buckminster Fuller that we can be architects of the future not victims of it. So I think that this opportunity to use art as a way to drive what could potentially happen with not only evolving storytelling but this idea that we could bring more inclusivity into the design of technology is really exciting.” To read more about the project, go to