This publication has previously reported on several promptonyms worth noting. When Elara Voss or Aris Thorne took over AI-generated content nearly a year ago, the speeding rate of AI development convinced this publication that the promptonym phenomenon would be solved relatively soon, and that we would, at least, be freed from infinite characters retaining the same name. Regrettably, it seems things have not come to pass. On the contrary, professional researchers and this publication have converged on the opinion that the promptonym phenomenon not only exists, but is actively getting worse.

For example, take the name Elena Vance to see whether we find all the classic signs of a promptonym, which, for viewers at home who may not be acquainted with the term, is a statistically favored name that large language models repeatedly and predictably generate, most visibly across various fictional characters. What are the signs of a promptonym? The most immediate tell is a cursory look at Goodreads or the Amazon Bookstore. Due to the flood of AI-generated literary works four years after the initial release of ChatGPT, these storefronts have become excellent samples of how AI-generated media names fictional characters. Blatantly, we do not even have to look at the actual character names in these works; the names of the supposed authors do most of the work for us. We find numerous examples of Aris Thornes, Elara Vosses, and now a couple of Alistair Finches thrown here, there, and everywhere. Does the name Elena Vance qualify under this threshold? Most certainly. We see an Elena Vance here, there, and here again.

The second sign of a promptonym is whether there are a suspicious number of “experts” using the name across various futuristic initiatives. For example, Dr. Aris Thorne, whose reach has only grown more pervasive, has now entered new fields of science, including, apparently, heaven. (We could not link the actual obituary; it links directly to a bot farm, perhaps predictably.) Beyond being featured in a fake obituary, he has since entered the ranks of “Genius Brain Experts,” appeared as a bathroom pipe technician in the online video game Roblox, and surfaced as an ER physician who watched a “thousand young men die on his tables” while opining on the execution of Christa Pike. Much like the execution itself, the morbidity rate on his tables implies something has gone terribly wrong, though we digress, since it is all entirely fabricated. The question is whether Elena Vance features as an expert in various degrees of seriousness, and she does. She reportedly turned around “Vance Global,” rose to the top of American cardiac surgery, and serves as a lead AI Fellow and senior editorial researcher. What a career. Therefore, Elena Vance passes both tests, i.e. she is undeniably a figment of the AI imagination.

An additional issue in detecting promptonyms is that some manage to overlap with the names of real people. While nobody names their child Sir Alistair Finch in the modern day, there are respected people named Marcus Chen, such as a researcher at Monash University and an Olympian born in 1991. On the other side of the promptonym scale, we find a Marcus Chen who is the owner of a popular AI News Instagram account, pitching himself as “The AI guy who tells you what’s happening in AI” (complete with the lighting emoji⚡⚡⚡). Then we see a Chen starring alongside the infamous Elara Voss in a story titled The 47th Floor, a Marcus Chen who has a thing or two to say about the hidden history of prompt engineering, and at least three different experts working in the generative AI space. Marcus Chen passes the promptonym test with flying colors, but buried somewhere beneath the pile of search results are real people. This makes the situation worse, as regrettably the collision of real identities with synthetic promptonyms pollutes search results and risks casting doubt on living professionals simply because an algorithm favored their name.

So, what are the academic publications converging on this conclusion? Researchers at the Samsung AI Center Warsaw (Michał Brzozowski et al.) had already documented in their preprint, The Ghost Couple: Correlated LLM Name Priors and Their Haunting of the Web and Academic Publishing, how models systematically mint recurring synthetic identities, such as Claude’s persistent pairing of Elena Vasquez and Marcus Chen, or Gemini’s affinity for Aris Thorne and Lena Petrova. Now, the journal Nature has published an investigation by Jackson Ryan titled “Elena, Aris, Marcus: AI-generated ‘ghosts’ are polluting the scientific literature”.

It is most unfortunate that the joke has ceased to be funny. The same statistical attractors that began as low-rent protagonists in self-published Kindle sci-fi novellas have, as detailed previously, migrated directly into the academic record, as hundreds of fraudulent papers across open repositories like Zenodo and ResearchGate now carry these synthetic names as credited co-authors. Apparently the real masterminds couldn’t take two minutes to check what the names actually meant. Worse still, because many of these bogus manuscripts have secured legitimate Digital Object Identifiers (DOIs), these phantom academics, Dr. Thorne, Dr. Vasquez, and Dr. Chen, are now quietly creeping into indexing engines, citation aggregators, and academic cross-references.

Perhaps it was inevitable that the lazy shortcut for churn-and-burn ebook mills metastasized into an automated hazard. People do use fake names for things more than ebook mills, after all. Still, perhaps we should be grateful, that the people behind said papers left a giant sign marked with this is an AI Generated Work, or at least an AI Generated author. Anyhow, it appears to be clear that the promptonym has migrated from laughs into our information commons, and until publishers, repositories, and AI labs implement verifiable human provenance, Dr. Thorne and his colleagues will continue to collect academic credentials they never earned.