The third wave of the survey of press office managers at German universities by Justus Henke (Institut für Hochschulforschung (HoF) an der Martin-Luther-Universität Halle-Wittenberg) und Matthias Begenat (Head of the Science Communication Department at CAIS) documents a rate of diffusion: While only 22% of university communicators used text generation tools in 2023, the figure had risen to 84% by 2025.
However, the study reveals a remarkable asymmetry between operational use and strategic integration: while 55% regularly use generative AI for text production, only 5% use the technology for target group analysis and 3% for impact measurement. Not a single press office uses AI to develop communication goals.
The newly developed KIWI-Matrix (“AI in Science Communication”) systematically highlights this discrepancy between operational and strategic AI use. The study also provides findings on the phenomenon of “shadow AI.” 69% of the communications managers surveyed state that generative AI is probably used more extensively in their department than is officially acknowledged. Employees resort to free services or privately financed accounts. This informal practice prevents transparent and reflective use.
Justus Henke / Matthias Begenat: Generative KI in der Hochschulkommunikation. Ergebnisse der 3. Welle – 2025 (HoF-Arbeitsbericht 135), Institut für Hochschulforschung (HoF) an der Martin-Luther-Universität, Halle-Wittenberg 2025, 62 S.
https://www.hof.uni-halle.de/web/dateien/pdf/ab135_WEB.pdf