Case Study | Week 6

Frankenstein AI

I shouldn’t write when I’m hangry.

It took me awhile to figure out just how Frankenstein AI works as a mechanism for collective storytelling. My first assumption, upon reading Lance Weiler’s blog post recalling the launch of the prototype at the Film Society of Lincoln Center, was that this was a collective retelling of the historic tale in a physical space. Lance (somewhat hypothetically) described what the project was, but not how it worked within a shared setting. I read and reread the descriptions of emotional catharsis among the participants and yet, remained uninformed about the storytelling mechanism:

“This energy was palpable in the room as the group quickly began to bond, and in the process, transformed the space into one of creativity and empathy…As the participants shared memories, emotions, fears and hopes, something unexpected happened. Humanity began to surface as philosophical conversations emerged.”

I grew exponentially more irritated at the fact that discussion questions and prompts weren’t shared on his blog. He spoke in theoretical vagaries, never really addressing how a new age Frankenstein story was/is amassed from a room full of people. How did he engage his audience? Where do data and code fit into this scenario? Get to the point, man!

It wasn’t until I listened to Robert Horton’s podcast, Frankenstein: A Monster Made By Many that I was able to discern how this immersive discussion was framed (but still not the means for conducting the data collection). Horton addressed the biographical points of Mary Shelley and how her radicalism led to such prolific and timeless writing. Her themes as author are very apparent: Frankenstein is the story of human’s desire to create and give birth to something. It is also the story of human’s failure to weigh unintended consequences. Horton addressed her imagery that has carried over into pop culture, through multiple movies and psychoanalytic archetypes. He discussed how the idea of the term “monster” is subjective, as the monster’s intellect grows throughout the story; the only thing that makes it seem monster-like is its physical appearance.

These were the themes I’d been missing in the original blog post. This is where the collective came into play over centuries of adaptations….but what about this current AI retelling? Why didn’t Lance just write about these themes in context? How did the audience participate in the storytelling? Did they talk in small groups? Share data online. And honestly, where does code fit in? I’m genuinely curious about storytelling through code, but neither the blog post nor the project website give a substantive technical explanation of how they compiled a modern composite of The Monster.

Which leads me to believe this is a junk project — a pie in the sky collaboration that has no structure and simply grifts off of 200 years of literary reviews, philosophers and remixes, topping off with talking points about AI and machine learning. In short, I don’t get the point or how this project actually happened (or is happening).

The Columbia DSL network has a private community forum where I can only guess that participants can engage with one another. It’s difficult for me to tell if this model of storytelling is at all effective. First thought: it’s a good try, but misses the mark. Horton’s insights were the most valuable aspect of this project. As a stand-alone lecture on the life of Mary Shelley and the evolution of Frankenstein, his thoughts are valuable and deeply fascinating.

However, I don’t understand where Lance’s AI project fits in? Is it even necessary? Did he make any new discoveries relating audience experience with artificial intelligence? Or is he just throwing around futurism terminology in an attempt to reinvent the wheel.

In short, perhaps not all projects need to happen. Sad to say, Frankenstein AI might be one of them.

Cover photo

Illustration of Mary Shelley by Irene Calvo Sanchez

Frankenstein movie poster