The TRAI project advances a comprehensive research agenda at the intersection of academic writing, critical thinking, and AI. Our objectives are structured to generate both theoretical innovation and practical educational impact.
RO1 – Mapping AI scholarship in thesis writing
We systematically analyse AI technologies currently associated with thesis writing and document the educational practices surrounding their use, including supervision models and institutional support.This objective establishes a conceptual framework for understanding how AI reshapes digital scholarship and redefines the intellectual demands of the thesis-writing process.
RO2 – Developing a corpus-based methodology for assessing critical thinking
We design a replicable methodology for assessing critical thinking and epistemic awareness in AI-augmented academic writing. This includes identifying linguistic indicators such as:
- Argumentation strategies
- Author positioning
- Use of hedges and boosters
- Levels of reflection and epistemic stance
The methodology integrates both traditional digital tools and AI-supported practices, reflecting the evolving landscape of academic literacy.
RO3 – Compiling a multilingual thesis corpus (TEZEC)
We build a balanced, multilingual corpus of Bachelor’s theses in:
- German
- Bulgarian
- Romanian
The corpus ensures disciplinary comparability and provides a foundation for cross-linguistic and cross-cultural analysis of thesis writing in the age of AI.
RO4 – Investigating the linguistic impact of digital technologies
Through cross-linguistic computational analysis, we examine how digital and AI technologies shape the language of thesis texts. This objective identifies genre-defining features within each national context and explores correlations between AI use and linguistic patterns.
RO5 – Developing research-informed teaching recommendations
Based on empirical findings, we produce pedagogical guidelines for thesis writing in AI-rich environments.
These recommendations aim to:
- Strengthen critical-thinking development
- Support responsible AI use
- Enhance epistemic awareness during the writing process
RO6 – Designing and implementing the multilingual platform (eTEZEC)
We design, build, and test an open-access digital platform that provides:
- Direct access to the TEZEC corpus
- Linguistic search and concordance tools
- Access to the critical-thinking assessment methodology
- Research-based teaching resources
The platform ensures sustainability and accessibility beyond the project lifecycle.
RO7 – Disseminating project results
We actively disseminate results through:
- Peer-reviewed publications
- Conference presentations
- Science communication initiatives
This ensures academic impact as well as broader societal engagement.
Planned key outcomes
The TRAI project produces both scientific advancements and immediately applicable educational tools.
‣ Research outcomes
- A replicable methodology for assessing critical thinking and epistemic awareness in AI-supported writing
- The multilingual thesis corpus (TEZEC) as a long-term research infrastructure
- Empirically grounded teaching recommendations for AI-integrated thesis writing
- A sustained body of scholarly publications and presentations
‣ Practical & educational outcomes
Multilingual educational platform (eTEZEC)
A centralized, open-access platform providing:
- Corpus search and concordance visualisation
- Access to analytical frameworks
- Research findings and pedagogical resources
MOOC course
A research-based online course focused on:
- Academic writing skills
- Critical thinking development
- AI literacy in thesis writing
Planned project milestones
The project progresses through a series of interconnected milestones:
- Completion of student and supervisor interviews
- Compilation and analysis of the first thesis corpus
- Development of the critical-thinking analytical framework
- Implementation of the MOOC intervention
- Second phase of interviews and corpus analysis
- Revision and refinement of the analytical framework
- Revision of the MOOC based on empirical findings
- Synthesis and integration of all data
