Hupmapper Redefines the AI Narrative by Scoring Human Original Intelligence

Hupside today opened the free public beta of Hupmapper, a new product that measures how much the ideas in a piece of writing stand out from what AI would produce for the same task. Rather than detecting AI use or judging the quality of the writing, Hupmapper measures what the person contributed on top of it and gives that contribution an Originality Intensity score.

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.

The Problem: Signal Collapse

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.

Source:

Tell Us What You Think

Do you have a review, update or anything you would like to add to this news story?

Leave your feedback
Your comment type
Submit

Sign in to keep reading

We're committed to providing free access to quality science. By registering and providing insight into your preferences you're joining a community of over 1m science interested individuals and help us to provide you with insightful content whilst keeping our service free.

or

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions.