# Old Ideas, Novel Problems: The Instability of LLM-Based Novelty Evaluation

arXiv: https://arxiv.org/abs/2610.02022 · HTML: https://mlsanity.com/paper/2610.02022/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 20 of its 38 references that are in the map.
- First posted on arXiv 1 Oct 2026.

## Builds on (most cited first, 20 of 20)

- [Large Language Models are not Fair Evaluators](https://mlsanity.com/paper/2305.17926/) (2023)
- [The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search](https://mlsanity.com/paper/2504.08066/) (2025)
- [Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers](https://mlsanity.com/paper/2409.04109/) (2024)
- [SciMON: Scientific Inspiration Machines Optimized for Novelty](https://mlsanity.com/paper/2305.14259/) (2023)
- [Towards End-to-End Automation of AI Research](https://mlsanity.com/paper/2606.15497/) (2026)
- [Is LLM-as-a-Judge Robust? Investigating Universal Adversarial Attacks on Zero-shot LLM Assessment](https://mlsanity.com/paper/2402.14016/) (2024)
- [Chain of Ideas: Revolutionizing Research Via Novel Idea Development with LLM Agents](https://mlsanity.com/paper/2410.13185/) (2024)
- [Learning to Plan & Reason for Evaluation with Thinking-LLM-as-a-Judge](https://mlsanity.com/paper/2501.18099/) (2025)
- [Nova: An Iterative Planning and Search Approach to Enhance Novelty and Diversity of LLM Generated Ideas](https://mlsanity.com/paper/2410.14255/) (2024)
- [Human-LLM Compound System for Scientific Ideation through Facet Recombination and Novelty Evaluation](https://mlsanity.com/paper/2409.14634/) (2024)
- [Evaluating and Enhancing Large Language Models for Novelty Assessment in Scholarly Publications](https://mlsanity.com/paper/2409.16605/) (2024)
- [Literature-Grounded Novelty Assessment of Scientific Ideas](https://mlsanity.com/paper/2506.22026/) (2025)
- [ScholarEval: Research Idea Evaluation Grounded in Literature](https://mlsanity.com/paper/2510.16234/) (2025)
- [Is this Idea Novel? An Automated Benchmark for Judgment of Research Ideas](https://mlsanity.com/paper/2603.10303/) (2026)
- [Measuring the Gap Between Human and LLM Research Ideas](https://mlsanity.com/paper/2607.01233/) (2026)
- [NovBench: Evaluating Large Language Models on Academic Paper Novelty Assessment](https://mlsanity.com/paper/2604.11543/) (2026)
- [Unlocking LLM Creativity in Science through Analogical Reasoning](https://mlsanity.com/paper/2605.11258/) (2026)
- [An Axiomatic Benchmark for Evaluation of Scientific Novelty Metrics](https://mlsanity.com/paper/2604.15145/) (2026)
- [It's Not What You Say, It's How You Say It: Evaluating LLM Responses to Expressions of Belief](https://mlsanity.com/paper/2607.18232/) (2026)
- [The Wording Effect: Quantifying Two-Way Drift in LLM Benchmark Performance](https://mlsanity.com/paper/2608.11694/) (2026)

## Abstract (arXiv)

AI-based ideation systems are often evaluated on the novelty of their ideas, and that judgment is increasingly delegated to large language models. Such judges are typically built ad hoc and validated, if at all, on human-authored papers rather than on the generated ideas they are meant to score. So, how do novelty judges perform? Not well. We present a systematic controlled study of novelty evaluation design choices. We build an evaluation set automatically, mining OpenReview for passages where reviewers explicitly affirm or dispute a paper's originality and keeping only submissions with unanimous agreement at the extremes of their research area; we pair these with ideas from a vanilla LLM generator. Across six judges, we find that seemingly small prompt design choices have large consequences; e.g., simply telling the judge that reviewers found one idea novel and the other not can change its verdict on more than half of the identical idea pairs it is shown, shifting pairwise accuracy by over 50 points and occasionally pushing it below chance. The same change helps one judge and hurts another. Retrieval and larger reasoning budgets help little, and two purpose-built novelty evaluators are outperformed by our cheapest prompted baseline. These results raise questions about reported novelty gains of automated ideation systems, and call for robust novelty evaluation methods.