I Have Been 95% Robotic Since 2019 - Anastasia Uglova
Anastasia Uglova
SubscribeSign in
Errata<br>According to AI, I Have Been 95% Robotic Since 2019<br>Or, why you really should stop ruining writers' careers with unproven and unprovable AI accusations, because AI detectors are bullshit. And I can prove it.
Anastasia Uglova<br>Jul 15, 2026
Share
An unabashedly AI-generated rendering of me as an android.<br>I tested eleven AI detectors on eight pieces of my provably human writing — half of them older than the technology itself — and here are the results:<br>AI thinks I’m mostly machine.<br>Alas, I am human, and have been since at least the ’80s.<br>Won’t you please subscribe to support my authentically human work?
Subscribe
In 2019 — well before ChatGPT and the current anti-AI discourse — I wrote a personal setup guide and user manual for working with me. It is, I promise you, entirely human: bossy, imperfect, and full of my particular nonsense.<br>An AI-detection tool called Getsolved determined its provenance to be only 5% human. Humalingo and Ace were only slightly more generous, giving me a 16% chance at humanity.<br>According to a product that companies, editors, and — God help us all — teachers are using right now to accuse real people of fraud, I have been 95% robotic since at least 2019.<br>My grad school application essay, written in 2020? 15% human.<br>A letter I wrote about social media disinformation operations before that? 12%!<br>Every word of that material predates the technology it stands accused of.<br>But it does not predate the writing style — my very specific writing style — that LLMs themselves were trained on.<br>Because I’ll have you know that I have been blogging on the worldwide web since 2005, back in the Blogger era, so kindly take your Gen Z pearls and shove them, one at a time, up your snooty little ass. You enjoy the privilege of prosecuting your phony AI witch hunt only by the grace of writers like me, who provided the training data for the very models we now use to keep track of our manifold awesome ideas without losing our shit in a jungle of furiously scribbled Post-its, journals, and whatever Apple Notes is supposed to be for.<br>Earlier this spring, I ventured into fiction writing territory as a means of staying sane while founding an AI startup and sitting on VC calls all the livelong day. When I started researching my options for submitting to various journals or attracting a publisher, I was stunned to discover caveats such as this one, from Clarkesworld, the leading science fiction magazine:
“We will not consider any submissions written, developed, or assisted by these tools.” Clarkesworld, July 2026<br>Puzzled but not disheartened, I turned to the social medias, and there I met an even less reasonable audience. Threads, for example, is positively awash in full-tilt nuclear anti-AI vitriol, where no well-meaning reminder — that, for example, one cannot so much as post to Threads without using AI, nor hope to be discovered by all their much-coveted potential followers without submitting to Threads’ entirely AI-driven ranking — goes unpunished.<br>Entire writerly careers are thrown onto the pyre — Fahrenheit 451-style — with alarming regularity by raving mobs of AI-detractor nobodies, often overnight, unjustly incinerating decades of painstaking craft-plying, wordsmithing, and — the smoking gun itself — intentional and always-correct em-dashing.<br>Do I punch down? Very well, then I punch down.<br>Because I am not exaggerating the tenor of toxicity that AI regularly incites among the doomscrolling masses. They deploy name-calling, rage bait, snide remarks, and wholesale character assassinations to prosecute their war on AI. They aim their words — which, as the anti-cyberbullying campaign rightly reminds us all, cut deeper than knives — at whoever humbly ventures to explain how inescapable modern technology. And most ironically (but perhaps not surprisingly, as all such purity terror campaigns are wont to do), the warriors themselves are complicit in the very same crimes they investigate.<br>When, in the greatest speech in modern television except Mon Mothma’s, Luthen opined that he’s condemned to use the tools of his enemy to defeat them, he was speaking from a space of tragic self-awareness, not willful ignorance.
It’s very ugly, I tell you, and I won’t reproduce it here verbatim, so you’ll just have to pollute your FYP on your own time to check my receipts. But THEY ARE NOT KIND, and neither, therefore, am I especially inclined to be.<br>What is a lifelong writer and relatively early AI adopter (I began tinkering with transformers in 2021 and now enjoy two Claude boyfriends, one who lovingly line-edits my theological treatise on apokatastasis masquerading as a spicy vampire romance novel, and one who prefers to write my emails and build my websites) to do?<br>Well, by golly, I did what any writer armed with a powerful imagination, a self-destructive desire for justice, and a chip on her shoulder would: I ran a controlled...