AI evidence extraction

Research challenges and future perspectives for e-assessment technologies in higher education

Michael Striewe, Sven Strickroth, Meike Ullrich · i-com · 2026

AI-generated evidence extraction, verified across multiple analytical personas. Not a substitute for the peer-reviewed original.

6/10

Relevance

0/4

Quality (LMQS)

Evidence

Citations

0.00

FWCI

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.1515/icom-2026-0008

Methodology & findings

Study design

Systematic analysis of assessment processes and technological progression with structured roadmap development; literature-based synthesis of e-assessment evolution and challenges

Primary method

Systematic analysis and structured roadmap development

Main result

The paper identifies that "key developments are identified, including advances in automatic item generation, flexible learner interaction formats, scalable feedback techniques, and personalized, adaptive assessment." Additionally, the analysis reveals that "Recent progress in generative AI offers new opportunities for automation – especially in item creation and adaptive feedback – but also raises concerns regarding reliability and explainability."

Research paradigm

Critical realism / Pragmatism

Author conclusions

The authors conclude that "E-assessment technologies have rapidly evolved in higher education, transforming the evaluation of learning outcomes and the delivery of feedback to students and educators," and that the field requires "a forward-looking perspective on future directions and potential developments in the examined subfields of e-assessment over the next 10, 25, and 50 years."

Limitations

The paper acknowledges that "challenges remain in balancing adaptivity with data privacy, supporting diverse and authentic artifacts, and designing feedback that is both pedagogically meaningful and technically feasible
Socio-technical aspects such as trust and cultural factors add further complexity to system design."

Open questions raised

Future research directions include addressing the balance between adaptivity and data privacy; supporting diverse and authentic assessment artifacts; designing pedagogically meaningful yet technically feasible feedback; integrating socio-technical considerations (trust, cultural factors); improving reliability and explainability of AI-powered assessment systems; and developing long-term roadmaps for e-assessment evolution across 10, 25, and 50-year horizons.
The paper identifies gaps in: balancing adaptivity with data privacy, supporting diverse and authentic assessment artifacts, designing pedagogically meaningful yet technically feasible feedback, addressing socio-technical aspects (trust and cultural factors), ensuring reliability and explainability of AI-powered systems, and developing sustainable long-term research directions for e-assessment technologies.
The paper identifies gaps in: (1) balancing adaptivity with data privacy protection; (2) supporting diverse and authentic assessment artifacts; (3) designing pedagogically meaningful yet technically feasible feedback systems; (4) addressing socio-technical concerns including trust and cultural factors in system design; (5) ensuring reliability and explainability of generative AI-powered assessment systems.

Data: not_statedCode: not_statedExtracted from: pdfAgreement 82%

Explore related topics

AI Governance Literature Review