The Oligarch Barely Steers Model Collapse in Multi-Model Ecosystems
arXiv cs.AIen
arXiv:2609.11146v1 Announce Type: new Abstract: AI-generated text is flowing back into the training corpora of the next generation of models. Recursive training on it drives model collapse, and recent work extends the setting to many models feeding one another -- but almost always with the market split evenly, while real generative AI is an oligopoly. Concentration raises two worries: fewer, more uniform sources may make collapse faster, and later models may be dragged toward the oligarch's output. We test both in controlled ecosystems: 13 open 1--4B models form natural ecosystems of 3 to 13 players, plus an injected probe that pushes the top share to 90%; each generation, every model's outp
This is a short summary published by AI Global Wire. The full article is owned and hosted by arXiv cs.AI — open it there to read it in full.
Read the full story at arXiv cs.AI- Forskning
- Företag
Related AI news
- Nvidia in talks to invest $10b in Anthropic IPO: sourcesTech in Asia · September 12, 2026
- Sequoia leads Mecka AI round at nearly $500m valuationTech in Asia · September 12, 2026
- CryptoL: Towards Scale Dominance and Physics Constraints Mitigation in Financial Multivariate Time Series ForecastingarXiv cs.AI · September 12, 2026
- Decoupling Readiness from Release for Tail-Aware Scheduling of Agentic LLM WorkflowsarXiv cs.AI · September 12, 2026
- Defining AI Agents: A Compendium of Criteria, Metrics, and BenchmarksarXiv cs.AI · September 12, 2026
- MOSAIC: Query-Aware Exploration Policy Adaptation for GraphRAGarXiv cs.AI · September 12, 2026