research
Three Years of Models Started Sounding More Alike
A preliminary longitudinal study found statistically lower diversity across open-ended model outputs over successive releases.
Summary
A preliminary longitudinal study found statistically lower diversity across open-ended model outputs over successive releases.
The study compares responses from three years of language-model releases on real open-ended prompts and the Alternate Uses Task. Sentence-embedding analysis finds a statistically significant decline in output diversity over time, suggesting convergence in creative substance even as model capability changes. The authors frame this as preliminary evidence: similarity metrics do not exhaust creativity, and the result does not prove that every model or human-AI workflow is becoming less original.
Why it matters
A preliminary longitudinal study found statistically lower diversity across open-ended model outputs over successive releases.
Limits and context
- The authors frame this as preliminary evidence: similarity metrics do not exhaust creativity, and the result does not prove that every model or human-AI workflow is becoming less original.
Key claims
A preliminary longitudinal study found statistically lower diversity across open-ended model outputs over successive releases.
Qualification: The authors frame this as preliminary evidence: similarity metrics do not exhaust creativity, and the result does not prove that every model or human-AI workflow is becoming less original.
Evidence: source-2026-08-21-005
Sources
- arXiv preprint 2608.19437arXiv · primary research
Corrections
No corrections have been recorded for this story.