AIght_
ToolsCompareLearnUniverseWorkflowsSignalHuman
Take the quiz
← All concepts

Concept

Model Collapse

What happens when models train on text written by other models — recursively.

Moon·Updated May 17, 2026
◈ Learn it interactively8 steps · ~6 min →

Where this idea lives

PREREQUISITESTOOLS THAT SHOW ITModel CollapseHow AI Models Are TrainedHow AI Models Are Trained — From random noise to a model that can reason — the actual pipelineScaling LawsScaling Laws — Why bigger keeps working — and the question of where it stops.Synthetic DataSynthetic Data — Models training on text other models wrote — and why this isn't always bad.ChatGPTChatGPTClaudeClaudeCommon misconception: Synthetic data is always toxic to training.Common misconception: We've already run out of training text.Common misconception: Model collapse is theoretical.
prereqsrelatedtoolsmisconceptions
shows up in:Environmental Science & ClimateHistory & HumanitiesJournalism & Media
You might think:Synthetic data is always toxic to training.We've already run out of training text.Model collapse is theoretical.

Common misconception

“Synthetic data ruins training, full stop.”

Carefully curated synthetic data — verified by humans or by another strong model — works fine and is used by every frontier lab. Model collapse describes what happens with uncurated generation-to-generation loops, where each generation amplifies the previous one's biases and narrows the distribution. The killer isn't synthetic data; it's unfiltered synthetic data.

Model collapse is the failure mode where each generation of a model, trained on the output of the previous generation, gradually loses variety. Rare patterns disappear; the distribution narrows. After a few generations, the model can produce confident, fluent text that has lost contact with the original data's diversity.

◉ INTERACTIVE

The Amazon rainforest is home to an extraordinary diversity of life. Jaguars stalk the undergrowth; macaws wheel overhead in flashes of scarlet and green. The river systems, ancient and labyrinthine, carry nutrients from the Andes to the Atlantic. Indigenous communities have mapped these systems for millennia, developing languages and practices that encode ecological knowledge unavailable in any database.

Vocabulary diversity score0.92

Original human-written corpus. Varied vocabulary, rare words, vivid specificity. The distribution is wide.

Why it happens

Models are probability distributions. Sampling from them tends to draw from the high-probability mass. Train on those samples, and the next model's distribution shifts toward the high-mass region. Repeat, and the tails — the rare correct cases, the weird-but-true facts — vanish.

Why it matters now

The web is filling up with AI-generated text. The next generation of models will train on it whether labs want to or not. Distinguishing human-written text from model-generated text is the live problem.

What labs do

  • Heavy filtering of training data — model detectors, source whitelists.
  • Synthetic data only where it's verified (math, code, structured tasks).
  • Curated human-written corpora (books, news archives, expert communities).
  • Watermarking attempts (still nascent).

What to read next

Synthetic data is the broader topic. Watermarking is one attempt at making AI-generated text detectable.

◈ Check yourself

You just read it — now see if it stuck. No score, no penalty for a wrong pick.

Question 1 of 3

What is model collapse?

Question 2 of 3

Why is model collapse described as a "photocopy of a photocopy" problem?

Question 3 of 3

As AI-generated text floods the web, what challenge does that create for training future models?

← Prerequisite

How AI Models Are Trained

Next →

Synthetic Data

by Moon · poet · engineer · third-year CS, Chandigarh University · Updated May 17, 2026

← Back to all conceptsBrowse tools →
intermediate
Read time4 min read
UpdatedMay 2026
Sources4

Read next

  1. How AI Models Are Trained →
  2. Synthetic Data →
  3. Scaling Laws →