A field guide to AI positions

Model collapse

Sources are over 18 months old.

Publication dates and source age

Sources counted: 2

Newest dated source: 2024-07-24

Assessed at this edition's evidence cutoff: 2026-09-15. 18-month boundary: 2025-03-15.

Publication age does not tell us whether a claim is still valid. Reading an old source again does not make its publication date newer. An update date does not establish that the passage we used was updated.

Open in the glossary Reading notes · Structured record

In plain language

Repeatedly training on earlier models' output can erase patterns from the original data. Other experiments avoided this degradation by retaining the original dataset while adding generated examples. [1] [2]

Reference this explanation or suggest a correction

Link to this explanation · Suggest a correction · How corrections work

Limits & distinctions

The result depends on how datasets are replaced or accumulated. It does not show that all synthetic data is harmful, or diagnose a chatbot's bad answer without examining its training. [1] [2]

Reference this explanation or suggest a correction

Link to this explanation · Suggest a correction · How corrections work

A fuller explanation

In studied training loops, each model supplies examples for the next. Errors can accumulate, and uncommon patterns can disappear from the later models' output. [1]

Reference this explanation or suggest a correction

Link to this explanation · Suggest a correction · How corrections work

How it relates to the map

Map context: this qualifies claims about generated data sustaining future model training. [1] [2]

Reference this explanation or suggest a correction

Link to this explanation · Suggest a correction · How corrections work

Share this page

https://theaiatlas.org/ideas/model-collapse/

Download a share image · Vector image

Image previews are summaries. Keep the page link so readers can check the evidence.

Sources and what we read

  1. 1. AI models collapse when trained on recursively generated data

    Source is over 18 months old.

    Publication dates and source age

    Sources counted: 1

    Newest dated source: 2024-07-24

    Assessed at this edition's evidence cutoff: 2026-09-15. 18-month boundary: 2025-03-15.

    Publication age does not tell us whether a claim is still valid. Reading an old source again does not make its publication date newer. An update date does not establish that the passage we used was updated.

    Read definition and recursive-training setup. Checked the 2025 correction to a mathematical symbol. Findings are not evidence that every synthetic-data method fails.

  2. 2. Is Model Collapse Inevitable? Breaking the Curse of Recursion by Accumulating Real and Synthetic Data

    Source is over 18 months old.

    Publication dates and source age

    Sources counted: 1

    Newest dated source: 2024-04-01

    Assessed at this edition's evidence cutoff: 2026-09-15. 18-month boundary: 2025-03-15.

    Publication age does not tell us whether a claim is still valid. Reading an old source again does not make its publication date newer. An update date does not establish that the passage we used was updated.

    Read abstract, introduction and language-model setup. Accumulating original and generated examples avoided collapse in tested settings. The original TinyStories text was itself synthetic.

Edition and machine-readable evidence

Content version 0.20.0. Evidence cutoff 2026-09-15; this does not mean every source was read on that day.

Pinned complete dataset · Complete evidence page · Agent consumption guide