# A Shared Taste for Model-Written Text: The Generator-by-Selector Matrices of "AI-AI Bias" Show No Detectable Own-Model Premium

arXiv: https://arxiv.org/abs/2610.00369 · HTML: https://mlsanity.com/paper/2610.00369/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 21 of its 25 references that are in the map.
- First posted on arXiv 30 Sep 2026.

## Builds on (most cited first, 21 of 21)

- [LLM Evaluators Recognize and Favor Their Own Generations](https://mlsanity.com/paper/2404.13076/) (2024)
- [Self-Preference Bias in LLM-as-a-Judge](https://mlsanity.com/paper/2410.21819/) (2024)
- [Preference Leakage: A Contamination Problem in LLM-as-a-judge](https://mlsanity.com/paper/2502.01534/) (2025)
- [Blinded by Generated Contexts: How Language Models Merge Generated and Retrieved Contexts When Knowledge Conflicts?](https://mlsanity.com/paper/2401.11911/) (2024)
- [Do LLM Evaluators Prefer Themselves for a Reason?](https://mlsanity.com/paper/2504.03846/) (2025)
- [Play Favorites: A Statistical Method to Measure Self-Bias in LLM-as-a-Judge](https://mlsanity.com/paper/2508.06709/) (2025)
- [Beyond the Surface: Measuring Self-Preference in LLM Judgments](https://mlsanity.com/paper/2506.02592/) (2025)
- [AI-AI Bias: large language models favor communications generated by large language models](https://mlsanity.com/paper/2407.12856/) (2024)
- [Self-Recognition in Language Models](https://mlsanity.com/paper/2407.06946/) (2024)
- [What Generative Search Engines Like and How to Optimize Web Content Cooperatively](https://mlsanity.com/paper/2510.11438/) (2025)
- [Self-Preference Bias in Rubric-Based Evaluation of Large Language Models](https://mlsanity.com/paper/2604.06996/) (2026)
- [AI Self-preferencing in Algorithmic Hiring: Empirical Evidence and Insights](https://mlsanity.com/paper/2509.00462/) (2025)
- [Quantifying and Mitigating Self-Preference Bias of LLM Judges](https://mlsanity.com/paper/2604.22891/) (2026)
- [Know Thyself? On the Incapability and Implications of AI Self-Recognition](https://mlsanity.com/paper/2510.03399/) (2025)
- [LLMs are Biased Evaluators But Not Biased for Retrieval Augmented Generation](https://mlsanity.com/paper/2410.20833/) (2024)
- [Are LLM Evaluators Really Narcissists? Sanity Checking Self-Preference Evaluations](https://mlsanity.com/paper/2601.22548/) (2026)
- [Self-Preference Is Weak or Absent in Verifiable Instruction-Following Revision: A Four-Model Test Under Genuine Authorship](https://mlsanity.com/paper/2606.20093/) (2026)
- [Evaluating Large Language Model Raters for German Open-Response Clinical Questions: A Physician-Annotated Benchmark Study of Agreement, Evaluator Bias, and Abstention](https://mlsanity.com/paper/2607.01103/) (2026)
- [On the impact of retrieved content representations in RAG Pipelines](https://mlsanity.com/paper/2605.30790/) (2026)
- [Self- and Other-Labels Induce Bidirectional Bias in LLM Judges](https://mlsanity.com/paper/2608.18091/) (2026)
- [Who Judges Matters: Measuring Family-Conditioned Preference in LLM-as-Judge Panels](https://mlsanity.com/paper/2609.17857/) (2026)

## Abstract (arXiv)

Laurito et al. (PNAS 2025) showed that large language models choosing between two descriptions of the same product, paper or film prefer the description written by a language model over the one written by a person, by a wide margin over what human judges do. Their design crosses five generators with the same five models as selectors, which permits a second question the paper does not headline: does a selector prefer text from its own model beyond what the generator and selector main effects predict? We rebuild the three 5x5 matrices from the per-item counts in the authors' public repository (21,828 valid trials; every cell matches the published value) and fit a two-way fixed-effects model with an own-model term gamma, tested by the exact permutation test over the 120 relabellings of the selectors. The premium is +0.013 on products (exact one-sided p = 0.24), -0.010 on paper abstracts (p = 0.74), +0.054 on films (p = 0.07) and +0.019 pooled (p = 0.14; 95% interval -0.008 to 0.046). The same-vendor term for the GPT-3.5 and GPT-4 pair is negative in all three datasets. Position bias moves single cells by up to 0.42 share points in either direction, and the own-model contrast is unchanged once order-driven items are removed. The design would have detected a premium of 0.05 with 82% (products), 88% (papers), 42% (films) and 97% (pooled) power; the minimum detectable effect at 80% power is 0.034 pooled. The absence is informative down to about 0.04 share points and silent below that. The 4x4 matrix of Tan et al. (ACL 2024) gives gamma = +0.148 at the smallest p its 24 relabellings allow, with a same-family term of the same size. The main result of Laurito et al. stands: models share a taste for model-written text, with GPT-4's descriptions chosen 77% to 95% of the time by every selector on products. What these data do not show is a model recognising and favouring its own prose.