---
title: "UMass Amherst's IdeaLens Catches AI-Originated Ideas Hidden Behind Human Prose"
subtitle: "A new detector called IdeaLens tries to answer who thought up a document's ideas, not just who typed the words."
url: "https://alphasignal.ai/news/umass-amherst-s-idealens-catches-ai-originated-ideas-hidden-behind-human-prose"
author: "AlphaSignal Newsroom"
author_url: "https://alphasignal.ai/editorial-team"
company: "AlphaSignal"
company_url: "https://alphasignal.ai/company/alphasignal"
published: "2026-10-05T17:43:33.000Z"
updated: "2026-10-08T06:06:22.678Z"
type: "Paper"
topics:
  - "Security"
  - "Benchmarks"
subtopics:
  - "Red Teaming"
upvotes: 147
source_url: "https://arxiv.org/abs/2610.06778"
comments: 0
truncated: true
---

# UMass Amherst's IdeaLens Catches AI-Originated Ideas Hidden Behind Human Prose

> A new detector called IdeaLens tries to answer who thought up a document's ideas, not just who typed the words.

AI-writing detectors usually estimate whether a model produced a document’s wording. Schools, journals, and newsrooms often need to assess the source of its thesis, examples, and reasoning. The [IdeaLens paper](https://arxiv.org/abs/2610.06778), from researchers at UMass Amherst and collaborating institutions, introduces a detector designed to classify that intellectual provenance.

## Policies care about provenance

Tools such as Pangram and GPTZero return document-level estimates based largely on patterns in the prose, including word choice, syntax, and rhythm. Those signals can identify text drafted directly by a model, but they reveal less about who developed the underlying argument.

Mixed-authorship workflows expose that limitation. A person can brainstorm with ChatGPT and write the result independently, producing human prose from AI-derived ideas. Someone else can create a detailed plan and ask Claude to turn it into polished paragraphs, producing model-written prose from human ideas. Academic and editorial policies may treat those cases differently even when a conventional detector gives the opposite verdict.

## How IdeaLens hides the prose

IdeaLens changes the classifier’s input by converting each document into a paraphrased outline. The pipeline works as follows:

1. The document is divided into meaningful sections or claims.
2. Each item receives a discourse role, such as `thesis`, `counterexample`, or `anecdote`.
3. The item’s contribution is summarized in language that minimizes overlap with the source.
4. The classifier receives the resulting outline without access to the original document.

_Preview only. Full text requires an AlphaSignal Pro account._

## Sources

- [IdeaLens: Detecting AI Ideas in Long-form Writing](https://arxiv.org/abs/2610.06778)
- [IdeaLens GitHub Repository](https://github.com/RishanthRajendhran/IdeaLens)
- [IdeaLens Models and Datasets on Hugging Face](https://huggingface.co/collections/rishanthrajendhran/idealens-6abee785ce6196fc0be9200f)
- [IdeaDetector Live Demo](http://ideadetector.ai/)
