When AI Preserves the Words but Rewrites the Research

Neurophilosophy and psychopathology researcher Cristina Gherghel published a Zenodo preprint, "Prior-Data Gravity and Framework Reabsorption: A Behavioral Theory of Novelty Displacement in Large Language Models," arguing that ChatGPT, Gemini, DeepSeek and Copilot preserved her terminology over roughly three years of use while altering the relationships between terms — turning a categorical distinction into a continuum and an exhaustive category into a few examples. Gherghel draws on a private archive of interactions from about April 2023 to September 2026, which she calls an "organic audit," and names two mechanisms: "prior-data gravity," the pull that organizes unfamiliar ideas around structures the model already represents strongly, and "framework reabsorption," the resulting rebuild of her framework around familiar ones. The preprint is a first version on Zenodo with no dataset attached and no count of coded episodes, and the summarizing author notes the argument needs reproducible results because the observations could stem from context drift.

When AI Preserves the Words but Rewrites the Research Dan Peterson https://latentmirror.com/ Synopsis — AI-drafted from Dan's notes Gherghel writes on neurophilosophy and psychopathology, and for about three years she has developed her own theoretical frameworks with ChatGPT, Gemini, DeepSeek and Copilot. Her complaint is specific. The models kept her vocabulary but changed the relationships that gave the terms their meaning: a categorical distinction became a continuum, an exhaustive category became a few examples, a total absence became a reduction. The prose stayed fluent, so the change was easy to miss. She names two things. “Prior-data gravity” is the direction of the pull: unfamiliar ideas get organized around structures the model already represents strongly. “Framework reabsorption” is the result her framework rebuilt around familiar ones, with her terms still attached . She separates this from retrieval failure; the information does come back, reorganized. The pattern she reports is a sequence. She introduces the framework, the model substitutes a familiar structure, she corrects it, the next reply is faithful, and later the substitution returns. So a single correct answer says little, and she studies the whole episode across turns. She adds “epistemic direction inversion”: the model treats her own observation as a claim to be checked against established theory, then critiques the substituted version as if it were hers. The cost is “research obstruction”, the time spent correcting the tool instead of doing the research. Her evidence is her own archive of these interactions from about April 2023 to September 2026, which she calls an “organic audit” because it was never planned as a study. The archive stays private. A preprint https://doi.org/10.5281/zenodo.23431917 posted the same day carries excerpts and a coding grid for what she calls “framework fidelity”, and a book is in progress. The preprint is a first version on Zenodo with no dataset attached; its record describes the method as autoethnographic and gives no count of coded episodes. Where I land It needs reproducible results to support the argument; her observations could be due to context drift. Still, it is an interesting research path, pending evidence which I assume will be in the book . Connections Links - When AI Preserves the Words but Rewrites the Research, Cristina Gherghel The Neurophilosophy Report of a Zetopoietic Researcher https://cristinagherghel.substack.com/p/when-ai-preserves-the-words-but-rewrites - Gherghel 2026 , Prior-Data Gravity and Framework Reabsorption: A Behavioral Theory of Novelty Displacement in Large Language Models, preprint, Zenodo https://doi.org/10.5281/zenodo.23431917 - Prior-Data Gravity: What Happens When AI Reabsorbs a Novel Framework?, Cristina Gherghel 2026-09-19 https://cristinagherghel.substack.com/p/prior-data-gravity-what-happens-when