Abstract

Researchers at the University for Continuing Education Krems, Austria asked for assistance with funding search and would appreciate automatically generated call recommendations. At the University for Continuing Education Krems we are committed to enhance our support for 370 researchers distributed across 18 departments and organized into three faculties. Each faculty has its own primary research focus, integrated into six transdisciplinary research areas. These areas encompass a wide range of subjects, from health and medicine to economics and globalization, and from education to arts and architecture.

The team for Grant Acquisition is dedicated to identify EU top-down calls tailored to the researcher’s profiles and topics and to foster a pro-active proposal development. With a team of 2,3 FTE grant officers, our challenge is to cover all different research disciplines at the university and to find a match with calls in the Horizon Europe funding program.

In a pilot study, we used AI (ChatGPT 4.0) to automatically create research profiles and keywords based on publication records to facilitate the funding search. Data from 13 researchers were collected from our institutional research database. Each dataset included the most recent five publications (full text or abstract), titles and descriptions of projects, and conference abstracts. We created a chatbot within ChatGPT 4.0 and fed it with 15 instructions based on self-learning material about creating worthwhile research profiles and relevant keywords. In this way, the chatbot drafted profiles of 13 researchers who were then asked to rate their own profile. 61% of researchers were satisfied with the generated keywords, and 46% of participants were satisfied with their research profile summaries.

In conclusion, enhancing pro-activity in the funding search through AI-generated research profiles and keywords significantly improves the efficiency of proposal development. According to data from the Austrian National Contact Point, this approach has the potential to increase success rates from 20% to 32%, thereby supporting the university's diverse research disciplines more effectively.

Learnings:
AI Application: Understand AI's role in generating research profiles for funding searches.
Effectiveness Evaluation: Assess AI-generated outputs for satisfaction and accuracy.
Interdisciplinary Support: Appreciate integrating diverse research areas for funding.
Data Utilization: Use institutional data to enhance AI processes.
Feedback Integration: Utilize researcher feedback to refine AI models.