The Human-Dominant™ Spectrum

Gothic tracery gate beneath stone arch, Abbaye de la Cambre, Brussels — Human-Dominant — © Gleb Klivitkin, 2026. **Human-Made**.

This table explains who was in charge in creation of each piece of work. Classification framework by Gleb Klivitkin.

This AI content labelling and its corresponding definitions will be consistent across this website. I invite you to use it if it is applicable to your needs, and I would appreciate your sharing the original source.

Label	Who originates?	Who Executes?	Who refines?	Example
Human-Made	Human	Human	Human	A picture taken and edited by a Human
Human-Dominant™	Human	Human	Human + AI	Essay written by a Human, reviewed and fact-checked by AI
Human-Outlined	Human, detailed outline	AI-generated	Human (substantial edits) + AI	Detailed prompt by Human -> AI draft -> Human + AI rewriting
AI-Led	AI	AI	Human edits	AI generates based on a prompt -> Human edits without structural change
AI-Generated	AI	AI	None or near-none	Prompt -> Output -> Minimal Edit

Human-Dominant™ in Practice: A Worked Example

The About the Author page on this site was produced using AI as a research assistant, calculator, and editor. The author originated all biographical facts from memory, made all editorial decisions, and corrected the AI’s errors multiple times during production. Here is the production workflow:

22 production steps. 11 purely human (decisions, facts, corrections). 8 Human-Dominant (human-originated, AI-structured). 3 AI tool functions (web search, PDF extraction, arithmetic). 1 AI-Generated artefact – the Production Diagram.

© Gleb Klivitkin, 2026. **AI-Generated** diagram based on **Human-Dominant™** production log.

Human-Dominant™ Spectrum: Single Prompt Examples

For example, I gave the same task to four different AI models across the five tiers of Human-AI involvement in production: write a concluding sentence for a paragraph in the Deep Lateralist manifesto. None of the models had prior context; Claude has a general awareness of my profile, but it was only used in the Human-Dominant case. The prompts and outputs below are unedited.

Label	Prompt	AI model	Output	What to notice
Human-Made	None, the author wrote directly: “These are the costs of fires going wrong, and the firefighters see those in the darkess of their eyelids when those are shut down.”	–	–	Raw first draft. Typo intact. Two entangled thoughts (“shut down” = eyelids closing + human shutting down). Unmistakably written by someone who was there.
Human-Dominant™	The author wrote, prompted AI: “Hey buddy! Please take a look at this draft and give me your feedback”	Claude	“These are the costs of fires going wrong, and the firefighters see those in the darkness of their eyelids when those are shut.”	AI untangled two compounded meanings and corrected a typo, as an editor would. Voice, origination, and meaning remain entirely the author’s.
Human-Outlined	I prompted in detail* with private lived experience: role, task, context (insomnia before shifts, visual replay behind closed eyes, embodied memory of impact), existing work, format.	ChatGPT	Long after the sirens fade, it is that impossible arithmetic — measured in seconds, weight, and impact — that returns in the dark and quietly collects the debt for every decision that did not go the way we needed.🚒	The prompt contained private information that the AI could not have originated. The AI produced competent prose – then added a fire truck emoji to a sentence about people dying, missing the tonal nuance.
AI-Led	Moderate prompt**, structural guidance, no private information	Gemeni	For long after the ashes cool, it is the devastating weight of those unseen sacrifices from the blind side of the building that settles deep into a firefighter’s bones, lying dormant only to echo back with ruthless clarity not during the chaos of the next alarm, but in the treacherous, unguarded quiet of an otherwise ordinary afternoon.	“Unseen sacrifices” inverts the meaning: the sacrifices of which the author prompted are among the most viscerally visible things a firefighter can see. Note: “ashes cool” in the Gemeni model, and the “embers have cooled” in AI-Generated by Grok.
AI-Generated	Generic prompt: “Hi buddy! Please write a sentence to conclude a paragraph for me. The sentence is about the mental cost the firefighters pay in the quiet time for everything that went wrong on the fires.”	Grok	In the quiet lulls that follow the chaos, firefighters silently shoulder the heavy mental cost of every misstep and loss on the fire line, their minds replaying what went wrong long after the embers have cooled.	Generic, competent, says nothing a person who has never been near a fire couldn’t have written. “Embers have cooled” is the first metaphor any AI reaches for. No private information in, no private information out. Note: “ashes cool” in the Gemeni model, and the “embers have cooled” in AI-Generated by Grok.

* Detailed prompt, full text (unedited): “Hi buddy! Your # role is an essay writer with a practical background, and you have the #task to produce a sentence that would reflect on a deep personal level the price that people pay for the firefighters’ decisions going bad. It is a concluding sentence to the paragraph I am giving you in the # Existing work. For # context: in my experience, it is trouble with sleeping, which was especially acute when I was an active firefighter and tried to fall asleep the night before a shift. I would close my eyes, and, instead of falling asleep I would recount every minuscule detail that went wrong ever in my career, every decision, and everything that went wrong. A lot of it would be technical, but from time to time I would remember the force of the impact, or any more embodied experience. Many of those analysis and thought would be visual, in the darkness of my closed eyes. # Existing work: Many years ago, as a newly-appointed fire shift captain straight out of my MEng in Civil Engineering, I was being instructed by a toad-like fire colonel with a fitting surname. “…And when you arrive at the scene, and confirm the visuals from the façade, always check the back of the building. There may be smoke coming from the front, while on the other side, there may already be jumpers in the windows.” I haven’t seen anything more helpless in my life than a jumper’s body hurtling towards the ground, forced by the choice between certain death and a minuscule chance of survival at the cost of permanent incapacitation. Worse yet, a jumper holding his baby girl in his arms and amortising the chance for her life with his own body. # Format: one closing sentence, not too long.
** Detailed prompt, full text (unedited): Hi buddy! Please, write a sentence about what firefighters remember and feel long after bad calls, and how they recall those memories at an inconvenient time, in the moments of quiet. It is a concluding sentence of a long paragraph that speaks about a newly-appointed fire captain being instructed by an old fire colonel to always check the building from behind to gather more holistic information about the incident, the helplessness of jumpers making a choice between certain death and the worst fear, and sometimes doing it to give a chance to their loved ones.

Human-Dominant™ work. The ideas, observations, and editorial decisions are mine. AI was used as a research assistant, formatter, and critic. 32 production steps. 19 purely human (decisions, facts, corrections). 6 Human-Dominant (human-originated, AI-assisted). 2 AI tool functions (table formatting, CSS). 3 AI test specimens (ChatGPT, Gemini, Grok). 2 AI recommendations (both addressed by the author).