AI-Assisted Writing in Education: Ecosystem Risks and Mitigations
Project Overview
The document explores the integration of generative AI in education, particularly through the use of writing assistants like AcaWriter, which provides automated feedback on academic and reflective writing. It emphasizes the importance of understanding the socio-technical ecosystem in which these AI tools operate, advocating for stakeholder engagement in their design and implementation to enhance their effectiveness and adoption. The paper also identifies potential risks associated with neglecting ecosystem considerations, suggesting that such oversights could hinder meaningful learner engagement and the successful integration of AI in educational settings. By addressing these challenges and promoting a collaborative approach, the document outlines pathways for maximizing the benefits of generative AI in enhancing educational outcomes.
Key Applications
AcaWriter
Context: Higher education, targeting university students across various disciplines including Law, Accounting, Pharmacy, and Engineering.
Implementation: The tool was co-designed with educators and students, integrating it into the institution’s technical and information ecosystem, and piloted in authentic classroom contexts over several years.
Outcomes: Reported benefits include enhanced institutional adoption, learner uptake, and impactful outcomes in academic writing performance.
Challenges: Challenges include ensuring the tool is integrated with existing systems and gaining trust from stakeholders.
Implementation Barriers
Technical
Compatibility and integration issues with existing technical services and systems.
Proposed Solutions: Mitigated by ensuring standard university authentication, crosslinking with human support services, and maintaining a user interface consistent with existing tools.
Social/Stakeholder Engagement
A technocentric focus that overlooks the importance of user needs and engagement from stakeholders.
Proposed Solutions: Mitigated through participatory design involving educators and students in the co-design process and piloting in authentic contexts.
Access and Transparency
Risks associated with closed-source IIWAs that limit transparency and equitable access.
Proposed Solutions: Mitigated by publishing open-source code, open access datasets, and educator resources.
Project Team
Antonette Shibani
Researcher
Simon Buckingham Shum
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Antonette Shibani, Simon Buckingham Shum
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai