The launch lands in the middle of a summer-long fight over how to police AI. New EU rules now push companies to watermark machine-generated text, and a crop of AI detectors promises to catch it after the fact. But the watermarks strip out with a copy-paste, and the detectors misfire, in one case flagging a writer's own mostly hand-written article as 100 percent machine after it circulated on Substack. Hupside argues that this fight is aimed at the wrong question. Whether a machine was involved tells you nothing about whether the person added anything original. That's what Hupmapper measures.
As more people lean on the same handful of AI tools, their output drifts toward the same baseline. The old shortcuts for judging quality, things like polish, structure, even credentials, stop telling you who is actually thinking. Hupside calls this signal collapse. In the workplace it shows up as strategy memos, board decks and customer emails that all read like the same author wrote them. In the classroom it shows up as essays and problem sets that are clean, confident and interchangeable. Watermarking does nothing about any of it. It just stamps the flood.
The Answer: Measure Originality, Not AI Use
Instead of asking whether AI touched a piece of work, Hupmapper asks a more useful question: how much original intelligence did the person add beyond what a baseline AI would have generated for the same task? It is a score you build, not an accusation you defend against.
"The whole industry is busy answering a question that stopped mattering," said Jonathan Aberman, CEO and co-founder of Hupside. "Everyone is using AI. Watermarking it, detecting it, arguing about it, none of that tells you whether a person brought anything of their own. That is the only thing worth measuring now, and it is what Hupmapper measures. We are not trying to catch anyone. We are giving people a way to prove the value they add."
How Hupmapper Works
Hupmapper is built around two connected capabilities, and it is the first product to combine them into a single, defensible measure of originality.
Learning Loop is a practice ground for expanding your own thinking. You bring real work, a board memo, a customer email, a strategy brief, a class essay, and iterate against a baseline-AI reference. Each pass returns a fresh score and a visible trajectory, so the work gets sharper and you build a skill you can carry into the next thing.
A score above zero means the content contains ideas or combinations of ideas that differ from what AI would be expected to produce. A zero does not mean the work was written by AI or that it lacks value. It means the concepts in that particular piece of work closely resemble the broad baseline of information and familiar patterns already available to large language models.
Proof of Work scores any piece of work against what baseline AI would produce on the same task, using a proprietary Hupside measurement model rather than a generic large-language-model rubric. Scores are consistent, reproducible and comparable across people, teams, business units, classrooms and time, giving leaders and educators an objective signal of where original thinking is, and is not, showing up.
Your Work Stays Inside Your Walls
The watermark debate has a privacy problem underneath it: the same techniques that mark AI text can, in principle, be used to trace who produced it. Hupmapper runs on Hupside-controlled infrastructure with no external LLM API calls at runtime, so artifacts, prompts and scores stay inside an organization's trust boundary. It is a posture built for regulated industries and other sensitive environments, and it pairs with Hupside's existing Hupchecker assessment for whole-organization originality coverage.
Experience It: Put a Number on Your Originality
The beta is free and open to anyone starting today. Additional features and paid availability for solopreneurs, organizations and institutions follow later this fall.
Hupmapper follows the earlier launch of Hupchecker and arrives alongside Aberman's book, The Originality Dividend: Why Human Original Intelligence Is the Most Valuable Asset in an Age of AI, which lays out the research behind Hupside's measurement model. Organizations and institutions interested in early paid access can request a conversation with the Hupside team.