In 1818, the year Frankenstein was published, the book featured a fictional man named Victor Frankenstein who robbed graves. In the nonfiction world of 1818 Europe, undertakers sold metal coffins as a way to ensure bodies stayed buried, otherwise the bodies were dug up and studied in the pursuit for knowledge. And staying buried was something only the rich could afford. For the poor, it meant that crowds fought at the gallows to protect the corpses of the hanged. That was the only way to ensure that the poor stayed buried.

This is a newsletter about AI.

Victor built his monster from the dead. Modern AI is built from dead cognitive labor: the archived writing, music, research, art, social media, medical records, and other data of millions of people over hundreds of years. All absorbed into proprietary models. Just like 1818, the poor are the ones least protected.

Welcome to The Necropolis Times. The lines between nonfiction and science fiction are still blurred. My writing fits in the blurry area … a fun area to exist in.

 

FRANKENSTEINING THE DISCUSSION

With a name like The Necropolis Times, this newsletter will explore topics some consider weird or uncomfortable.

Mary Shelley wrote one of the founding texts on weird, macabre, disturbing … and a commentary on the responsibility of creators: Frankenstein; Or, the Modern Prometheus.

For Shelley’s book, the name of “the monster” has slipped from Victor to his creation. Many people mistakenly call the monster Frankenstein.

Two centuries later there are different interpretations about who the monster was in that book. Was the monster Victor Frankenstein? Was the monster the creature Victor created?

 

HOW DOES AI RELATE TO FRANKENSTEIN?

Aaron Hertzmann wrote that SkyNet (from the Terminator franchse) is Frankenstein's monster, with neural networks as the Promethean spark instead of galvanism. In a footnote, Hertzmann notes that Frankenstein is a cautionary tale about the quest for knowledge in general.

But whether we are discussing Skynet or Frankenstein’s monster, the story trajectory stays the same: The creator builds a machine, the machine wakes up, the machine attacks the creator, the creator learns his lesson too late.

Victor built his monster from the dead. He robbed graves and charnel houses for the parts. Nobody that Victor took the parts from agreed to any of it. Those were not anonymous parts. David McNally discussed this in Monsters of the Market: Zombies, Vampires, and Global Capitalism. Eighteenth-century London crowds fought anatomists at the gallows for the bodies of the hanged, because surgeons and resurrectionists were understood as one apparatus for degrading the poor in death as in life. The first edition of Frankenstein appeared in 1818, a year that also saw metal coffins marketed in Europe as a way to thwart grave robbers. McNally points out that the poor, of course, could not afford the luxuries of armed guards to protect their graves or metal coffins.

McNally discusses Shelley because the story of Frankenstein highlights how science and science fiction meet in an uncomfortable way, because the lines of fiction and nonfiction start to blur.

Modern AI systems are assembled in a similar fashion to Frankenstein, except instead of robbing body parts, the accumulation involves dead cognitive labor: the objectified traces of human intellectual work such as archived texts, images, code, and communicative recordings which are stripped from their creators and reconstituted within proprietary model parameters.

The people who built training corpora for data sets followed a similar pattern to the Eighteenth-century London grave robbers. Yes, they procure data instead of body parts. But just like the Eighteenth-century grave robbers, in modern times it is the poor who are most at risk. Not of body parts being taken, but land and water and electricity being taken by data centers in the quest to accumulate and process more data (Patterson and Haynes 2025; DiGangi 2024; Chari 2025).

The grave robbers, and the data center supporters, are doing it in the name of science with apparently little regard for who gets hurt in the process.

Who gets hurt goes beyond just building and maintaining the data centers. It includes the raw gathering of data via violations of privacy, ownership, intellectual property, individual creativity.

Will Orr and Kate Crawford interviewed dataset creators and found the procurement question was handled by declining to ask it. One creator conceded that consent and privacy are real concerns for scraped data, then admitted the team had never thought much about them, and that thinking about them might have ended the project before it started. Maybe it could have been cognitive dissonance, except they apparently didn’t put enough thought into it for cognitive dissonance to occur.

Artificial intelligence is created by people who didn’t want to put too much thought into how it was created. That’s where current AI systems overlap with Frankenstein. And that uncomfortable macabre area is where this newsletter aims to exist, by exploring the uncomfortable, the impolite, the critical thinking space that occupies both science and science fiction.

 

THE PUBLICATION SCHEDULE IS IRREGULAR. THE TOPICS WILL BE IRREGULAR.

I write other things. I teach. I have a family.

I started this newsletter in 2026 as an informal extension of my peer-reviewed articles. I love science fiction books and movies. I play video games but not as often as I used to. Whether the book or movie is considered space opera, science fiction, campy sci-fi, hard science fiction, or some other genre is a discussion for a different publication.

This newsletter, for me, is a place to explore ideas that fall outside the mainstream AI discourse, and occasionally provide social theory interpretations. GenAI, true AI, machine learning, and data centers will all be touched on from time to time.

Sometimes it’s just fun, so don’t take it too seriously.

Sometimes I might write a book review. Other times a more in-depth commentary on a news story. Occasionally I’ll share a commentary about a movie or TV show that illustrates something interesting about culture and AI.

Thanks for reading. This newsletter is free. Don’t forget to subscribe if you have not already done so.

DEFINITIONS FOR TERMS I WILL USE:

Dead Cognitive Labor: the objectified traces of human intellectual work such as archived texts, images, code, and communicative recordings which are stripped from their creators and reconstituted within proprietary model parameters (e.g., a human wrote a research article, the text is scraped into a training corpus, and then the paraphrased research article is sold back as ‘original’ work with no credit and no citation).

Cyberzombies: a proprietary system that animates dead cognitive labor as if it were autonomous cognition while remaining structurally governed by capital. Cyberzombies are not labor themselves. They are capital’s instrument, not a worker with agency. They are dead cognitive labor hardened into capital, confronting living cognitive labor as an independent power and surviving by feeding on continued human labor (e.g., a studio spends decades developing a house animation style, that style is absorbed into a model, and then that model creates images on demand while the living artists who created that specific animation style go unnamed and unpaid). In other words, cyberzombies are fed on the output of human brains.

Cognitive Rift: the divide between the communities that supply the material, intellectual, physical, and emotional resources required to power AI and the institutions that accumulate the resulting profit and control (e.g., a data center in a West Virginia county processes workloads for clients the county never sees, and the residents see their water and electric bills go up).

The necropolis of capital: the infrastructure itself, where the scraped dead cognitive labor of centuries of human work is stored and reanimated on demand.

 

Those are the basic terms, drawn from my peer-reviewed articles (listed below).

The cyberzombies and ghosts are the result of dead cognitive labor. The monsters are varied.

 

 

 

REFERENCES:

Cameron, James, dir. 1984. The Terminator. Los Angeles, CA: Orion Pictures.

Chari, Maya. 2025. “Are Data Centers Depleting the Southwest’s Water and Energy Resources?” APM Research Lab, February 27. Retrieved November 19, 2025 (https://www.apmresearchlab.org/10x/datacenters-resource).

DiGangi, Diana. 2024. “Duke Energy Signs Agreements for 2 GW of Data Centers in Past Month: CFO.” Utility Dive, November 11. Retrieved November 7, 2025 (https://www.utilitydive.com/news/duke-energydata-centers-load-growth-q3-earnings/732491/).

Hertzmann, Aaron. 2018. "Can Computers Create Art?" Arts 7(2), Page 15. May 8. Retrieved August 10, 2026. doi:10.3390/arts7020018.

McNally, David. 2012. Monsters of the Market: Zombies, Vampires, and Global Capitalism. Chicago, IL: Haymarket Books, Pages 12, 57.

Orr, Will, and Kate Crawford. 2023. "The Social Construction of Datasets." SocArXiv. doi:10.31235/osf.io/8c9uh_v1. November 7. Page 24. Accessed August 15, 2026. PrePrint.

Patterson, Destinee, and Carly Haynes. 2025. “Residents Push Back Against Proposed Data Center in Edgecombe County.” WRAL, September 8. Retrieved November 7, 2025 (https://www.wral.com/news/local/data-centers-edgecombe-county-residents-mixed-feelings-tarboro-sept-2025/).

Shelley, Mary. [1818] 2008. Frankenstein; or, The Modern Prometheus. Project Gutenberg. Retrieved August 10, 2026 (https://www.gutenberg.org/ebooks/84)

Van Pelt, Craig. 2026. “Cyberzombies and the Afterlife of Thought: Dead Labor, AI, and the Necropolis of Capital.” Fast Capitalism. 23(1): Article 4. https://mavmatrix.uta.edu/fastcapitalism/vol23/iss1/4/

Van Pelt, Craig. 2026. “Dead Cognitive Labor and the Tempo of Machine-Time: AI, Infrastructural Capture, and the Privatization of the General Intellect.” Sociation, 26 (2), 19-34. https://www.researchgate.net/publication/412085151_Dead_Cognitive_Labor_and_the_Tempo_of_Machine-Time_AI_Infrastructural_Capture_and_the_Privatization_of_the_General_Intellect