Predictive AI for Socially-Aware and Legally-Compliant Research Policy in CRIS Systems
Author(s)
University of Szeged
FernUniversität in Hagen
Issue Date
May 22, 2026
Publisher
euroCRIS
Type
Conference Paper
Abstract
This contribution investigates the use of predictive AI models to develop socially-aware and evidence-based research policy within CRIS systems. Predictive AI serves not only to forecast future research developments but also as a tool to simulate complex socio-technical interactions where policy decisions, scientific practices, legal frameworks, and societal expectations intersect. The aim is to enable decision-makers to compare different policy scenarios, anticipate potential impacts, and make strategic decisions based on solid evidence. As humorously noted, if a model prediction “fails,” it can still provide a learning opportunity—predictions have two sides, and their uncertainties offer insights for governance strategies.
Methodologically, the approach combines multiple AI techniques within the CRIS architecture. Machine learning analyzes historical trends and impact patterns; causal models identify key policy influence factors; agent‑based simulations capture individual and institutional behavioral changes; and Natural Language Processing analyzes policy documents, regulatory requirements, and qualitative reports. By linking internal CRIS data with external sources — such as Open Access monitoring, funding statistics, compliance data, or social indicators — scenarios can be simulated to test hypothetical policy interventions, such as changes in evaluation criteria, Open Access mandates, funding logic, or governance structures.
Description
Presentation delivered within CRIS2026 session "AI-Enhanced CRIS for Decision Support, Prediction, and Policy".-- Includes extended abstract and video from Gizem Gültekin's talk
URI
https://dspacecris.eurocris.org/handle/11366/10589
File(s)![Thumbnail Image]()
![Thumbnail Image]()
Name
CRIS2026_paper-7_Gültekin-Azeroual_Predictive-AI_extended-abstract.pdf
Description
Extended abstract
Size
1.04 MB
Format
Adobe PDF
Checksum (MD5)
a01ee553c91d4bacdc2cc4112c348434
Name
CRIS2026-GGültekin_video_Predictive-AI.mp4
Description
Video
Size
32.28 MB
Format
MP4
Checksum (MD5)
daa836bb85f2ae6372af5582464612a6
Conference(s)