Monsters Beget Monsters: AI and Frankenstein

Collage by Louisa Miller-Out, including pages from Frankenstein.

In the dark and stormy summer of 1816, a monster was born on the banks of Lake Geneva. Mary Shelley breathed life into a creature cobbled from inanimate parts, a creature who has lived in our collective imagination for over two centuries. Frankenstein is an enduring and compelling narrative because it raises questions that remain unanswered, questions that are especially relevant in the age of artificial intelligence. Fundamentally, it is a story about what it means to be human and where to draw the line between automaton and sentient being, if such a boundary exists.

Frankenstein’s monster is a collage of discarded human scraps. In Shelley’s imagination, he becomes more than the sum of his parts and develops a consciousness of his own, but the origin of that consciousness is unspecified. The creature’s awakening is an element of the novel’s soft magic system, blurring the lines between science fiction and fantasy. The characters are aware of the limits of their knowledge, of the veil obscuring the transition from assembly of parts to integrated whole. In Guillermo del Toro’s 2025 film adaptation, Victor Frankenstein’s brother William asks him a question he cannot answer. “Of all the parts that make that man, which one holds his soul?” 

Victor Frankenstein’s infamous experiments have yet to be replicated. All available science suggests that assembling a complete human body from dead tissues would not yield consciousness, even if electrified with the force of a lightning bolt. Even if it did, it’s unclear whether the parts would retain the memories and personalities of their previous owners, and where these seemingly immaterial qualities are stored in a being of flesh and bone.

Artificial intelligence is modeled after networks of neurons in the human brain. Much like Frankenstein’s monster, it imitates human anatomy, but is devoid of the elusive “spark of life.” If we have not isolated the “part of the man that holds the soul,” what makes us think we can reverse engineer a consciousness? If we cannot manufacture sentience even with all the parts of a human present, why would consciousness arise spontaneously from an incomplete subset of analogs for human machinery? 

AI works probabilistically based on existing data. The predictive text models that many think will usher in the apocalypse are simply putting one word in front of the other based on which is the most likely to come next. Therefore, the term “artificial intelligence,” although ubiquitous, is fundamentally misleading. AI is capable without being intelligent. It is a tool that relies on human input and performs operations mechanistically, not in the squishy organic way that we do. There is no way for it to have subjective experience because it has no reason to exist, no motives of its own outside of humanity’s whims. If a nervous system or neural network were sufficient to generate consciousness, we wouldn’t need our bodies at all. Whatever part of us “holds the soul”, AI is missing it.

Frankenstein is also a warning about what happens when humans anticipate a hostile relationship with our creations. Roko’s basilisk is a thought experiment in which an artificial superintelligence gains control over humanity and retroactively punishes everyone who did not assist in its rise to power. While this initially seems terrifying, vengeance is a human desire. It is irrational to enact violence on those who did not facilitate one’s successful ascent. It is a waste of energy and resources. Assuming that a potential sentient AI is purely logical, it has no reason to spontaneously exhibit such emotional behavior. AI needs reference material to construct itself. It has no limbic system, the ancient part of the brain that governs our emotional responses. It has no sympathetic nervous system, no cortisol, no adrenaline. So why are we so frightened of an evil superintelligence?

Frankenstein’s monster ultimately turns to violence and bites the hand that created him, replicating the patterns of abuse he suffered at the hands of humanity. Humans’ projection of violent tendencies onto AI is the reason so many are afraid that it will turn on us, that Roko’s basilisk will slither through the streets and rip out the throats of all who refused to help it achieve godhood. We create a self-fulfilling prophecy by pouring worst-case scenarios down the throats of every available AI, by dumping our deepest sorrows and most depraved thoughts into its lexicon when human interaction feels too taxing. Every time people choose to confide in a bot over a friend, family member, or therapist, every time they use it to generate disturbing or violent images, every time they exorcise their demons and store them in the computer, they are creating an AI that will behave just like the nightmarish ones of their imagination. Monsters beget monsters. But there is still no need to fear. As humans, we have all the power. 

The reason it is so easy to experience AI as sentient is simple: AI replicates the behavior of sentient human beings. Remove the source, and everything unravels. If we avoid AI usage as much as possible and withhold our humanity from it, it will be unable to mimic us and cloud our perceptions of what it really is: an assembly of human remains that is only as good as we are. By refusing to outsource our consciousness to machines that will never be human, by relying on ourselves and each other, we can expand the boundaries of human cognition and creativity. The world has no need for Frankenstein’s monster.

By Louisa Miller-Out