Concern 496 · Working draft · editorial review in progress
Reference-corpus decay
Reference-corpus decay occurs when unreliable generated material becomes part of the information used to train or inform later systems.
Understanding the concern
A corpus is a collection of material used as a reference or training resource. If errors from one system are copied into that collection, another system may learn or repeat them. It resembles a photocopy of a photocopy when each generation loses useful detail, although real training pipelines can also introduce fresh, checked information.
Research on repeated training with generated data has demonstrated problems under specified conditions. It does not show that all synthetic data is harmful. Carefully selected synthetic examples can be useful. The practical issue is maintaining data quality, variety and a traceable connection to reliable observations rather than assuming that fluent text is sound evidence.
An example
Several systems repeat an invented fact until it appears common in material collected for training.
Illustrative unless identified as a documented case.
Words in plain English
- Training
- The process of using examples to shape how an AI model behaves. Learning patterns from material does not establish that the material is true or fair.
- Corpus
- A collection of material, often text, used to train or study a system. Corpora is the plural.
- Synthetic
- Made by a computer or another artificial process rather than directly recorded from the event being described.
Evidence and its limits
The study tests recursive generated-data training under stated assumptions. It does not prove that the entire web or every future model will deteriorate.
Shumailov and colleagues · Nature 2024 · Research study
AI models collapse when trained on recursively generated data ↗ (opens in a new tab)Examines repeated generated-data training under stated assumptions. It does not show all synthetic data is harmful or that collapse is inevitable in every pipeline.
Broader frameworks and background
These resources provide context. They do not establish every claim in this concern.
- OECD AI Incidents and Hazards Monitor (opens in a new tab)
incidents
- AI Incident Database (opens in a new tab)
incidents
- International humanitarian law (IHL)
framework
The body of law that governs the conduct of armed conflict and the protection of people who are not fighting, which applies to AI used in or for conflict.
A question to discuss
How can a training collection preserve original evidence and avoid treating repetition as confirmation?
Keep track of different views, what needs checking, and any next step in your own discussion notes.
Open discussion notes →Principles for the conversation
Before you finish · reflect together
Virtues and principles to consider
Virtues are qualities we practice in how we treat people and make choices. They are not scores for judging someone. Use these ideas to explore what a thoughtful response could look like.
Truthfulness
Truthfulness means making claims that match the evidence and being clear about what we do not know. It includes correcting mistakes rather than protecting a polished story.
Check sources, label guesses, and distinguish a trial result from a proven benefit. If AI helped produce a claim, someone still needs to take responsibility for checking it.
Ask together: What can we honestly say, and what still needs checking?
Meaningful conversation
Conversation means making room to understand and be understood. It is more than winning a debate or collecting quick agreement. People can explore a question together without starting from the same view.
Use a concrete example, invite different experiences, and ask what might change each person's mind. Let concerns and hopes sit together long enough to reveal choices that a simple yes-or-no argument would miss.
Ask together: What can we learn from each other before deciding?
Unity
Unity means working toward shared purposes while making room for real differences. It does not require everyone to have the same experience, beliefs, or opinion.
Look for a goal people can share, then explore disagreements about how to reach it. Make sure quieter or less powerful voices can influence the result.
Ask together: What purpose can we share while respecting our differences?
Editorial ratings
These are editorial judgments from the register, not measured probabilities or predictions about your situation.
About the register and scoring →More ways to explore
Search for more writing
For more specific or recent material, open a search below or ask for a web summary. AI-generated summaries can contain mistakes; open the original articles to check them.