European universities write for multiple audiences at once. AI tools are flattening these registers into a single promotional voice. The scholarly tone is giving way to something closer to marketing.
Three findings from the Academic sector corpus. Each carries statistical evidence from the full PRISM™ analysis.
Citation density fell. The academic habit of attributing ideas, referencing precedent, and situating claims within a literature is diminishing.
Oxford, ETH Zurich, and Bocconi converging despite different scholarly traditions, languages, and national academic cultures.
Generic phrase density rose. "World-leading", "cutting-edge", "transformative". The promotional register is replacing the scholarly one.
Current PLI scores versus 2015–2019 baseline. Sorted by drift magnitude.
Current PLI score versus 2015–2019 baseline. Scores with fewer than 10 documents in the current window are excluded. ⓘ Ordering rationale
| Organisation | Current PLI | Baseline PLI | Coverage |
|---|---|---|---|
| Oxford | 68.4 | 78.2 | |
| Cambridge | 67.2 | 77.8 | |
| ETH Zurich | 66.8 | 76.4 | |
| LSE | 65.1 | 74.2 | |
| TU Munich | 64.9 | 73.5 | |
| Bocconi | 64.5 | 73.8 | |
| Sciences Po | 63.8 | 72.4 | |
| KU Leuven | 63.2 | 72.1 | |
| Sorbonne | 62.4 | 71.8 | |
| Bologna | 61.8 | 70.4 |
Organisations are ranked by PLI score. The confidence badge reflects the number of scored documents in the current window: High (≥ 30 documents), Medium (10–29), Low (fewer than 10). Low-confidence scores are included in the ranking and should be read alongside the document count.
Sector median with interquartile range (shaded). Corpus median shown as dashed reference line.
Ten PRISM™ dimensions. Baseline (dashed) versus current (solid). Sorted by drift in the table.
| Dimension | Baseline | Current | Δ | |
|---|---|---|---|---|
| Citation Density | 84.0 | 55.0 | -29.0 | |
| Cultural Markers | 74.0 | 58.0 | -16.0 | |
| Idiom Density | 70.0 | 56.0 | -14.0 | |
| Perplexity | 76.0 | 68.0 | -8.0 | |
| Burstiness | 72.0 | 64.0 | -8.0 | |
| Lexical Diversity | 78.0 | 70.0 | -8.0 | |
| Jargon Load | 78.0 | 72.0 | -6.0 | |
| Hedging Language | 68.0 | 62.0 | -6.0 | |
| Readability | 62.0 | 58.0 | -4.0 | |
| Generic Phrases | 60.0 | 71.0 | +11.0 |
Cross-sector linguistic similarity. Higher scores mean Academic is beginning to sound like other sectors.
Quality metrics for the Academic sector corpus used in this analysis.