AI & writing assistance

Heiko Pääbo (University of Tartu): AI as grading assistant (writing assistance) 

Over the past eighteen months, I’ve integrated generative AI into my assessment practices across various MA-level courses with student cohorts ranging from 15 to 30. This experience has revealed several practical ways AI can support key teaching objectives while addressing time constraints often limiting feedback quality.

My own experience has shown that OpenAI’s ChatGPT is a powerful tool for this. Copilot can be used, but based on my experience, ChatGPT is more flexible for in-depth analyses. Academic text evaluation tools such as CoGrader frequently fail to effectively analyse content and offer insightful feedback.

My primary application has been using AI to create initial drafts of student feedback. Instead of beginning each assessment from scratch, I prompt the AI to identify fundamental issues in structure, evidence use, and argument development. This creates a foundation I can then refine with my subject expertise and knowledge of individual students’ progress.

This approach has substantially improved both my feedback quality and consistency. Students receive more detailed comments addressing both technical writing issues and conceptual understanding. The structured format ensures all submissions receive thorough attention to essential scholarly elements while still allowing for personalised guidance tailored to each student’s work.

A crucial aspect of this integration is maintaining careful oversight of AI-generated content. While effective at recognising patterns and identifying common issues, current AI tools sometimes misinterpret disciplinary conventions or specific assignment requirements. My academic judgment remains essential in ensuring feedback aligns with module objectives and genuinely supports student development.

Rather than replacing academic expertise, AI functions as a complementary tool that manages routine assessment components. This creates space for more meaningful engagement with the sophisticated analytical skills and subject mastery central to postgraduate education. The approach preserves essential evaluative dimensions while significantly improving efficiency in feedback delivery.

This balanced implementation addresses both educational imperatives and practical workload considerations, offering a sustainable model for maintaining high-quality assessment practices despite increasing pressures on academic time and resources.