Transparency in academic work is paramount, particularly regarding using AI tools in educational settings. By fostering an environment of transparency, we can build trust, ensure integrity, and maximise the benefits of AI in universities while mitigating potential risks. This transparency involves an open declaration of AI tool usage, a clear distinction between AI and human contributions, insight into AI functionality, and accessible and detailed information on AI-related policies from universities.
The importance of transparency in AI usage extends across multiple dimensions of education. From an ethical standpoint, it promotes the responsible use of AI tools while addressing potential biases and limitations. Open disclosure fosters trust among stakeholders in the academic community while ensuring accountability for students and educators. Furthermore, transparency develops students’ understanding of AI capabilities and limitations, enhancing their critical thinking skills and enabling educators to evaluate learning more effectively.
However, using AI tools in academic writing raises significant questions about originality and authenticity. Recent cases have highlighted the need for clear guidelines on AI usage in educational work. Achieving true transparency remains challenging due to difficulties in detecting AI use, rapid technological evolution, and limitations of existing detection methods. Addressing these challenges requires ongoing dialogue about AI in academia, critical evaluation of AI tools, robust integration frameworks, and enhanced AI literacy.
The Grammarly Authorship Model
Grammarly’s recently introduced Grammarly Authorship, a transparency model, represents a significant advancement in addressing AI-assisted writing challenges in academic settings. Unlike the standard Grammarly product, which focuses on grammar and style suggestions, the Authorship model provides comprehensive insights into the writing process, including AI, and promises to maintain transparency and integrity in academic work (although I need to do more testing).
The model automatically categorises text as it enters a document, distinguishing between human-typed text, AI-generated content, AI-assisted edits, and content from various sources. This granular categorisation and comprehensive reporting features provide a clear picture of document creation. The reports include document analytics, writing time statistics, and colour-coded text indicating origin, offering valuable insights into the writing process.
The Authorship model offers significant benefits for students and educators. It allows students to demonstrate work authenticity and protects the learning process. The feature facilitates constructive discussions between students and faculty about AI use in assignments, moving away from after-the-fact detection toward a more collaborative approach. Educators can utilise these reports to identify areas for improvement and provide personalised instruction.
Currently available as a beta feature and only in Google Docs, Grammarly Authorship requires users to install the browser extension, enable tracking at document start, and grant clipboard access. Privacy considerations are paramount, with users maintaining complete control over feature enablement and data encryption until report generation.
Grammarly plans to expand the Authorship feature by early 2025, introducing proactive nudges for citing external sources. This development demonstrates a commitment to evolving tools that meet the academic community’s changing needs in the AI era. As AI continues integrating into various aspects of academic work, tools like Grammarly Authorship will be crucial in fostering a transparent, ethical, and productive learning environment.
AI Declaration
In creating this work on “transparency in AI usage and the Grammarly Authorship model”, I utilised several AI tools:
- NotebookLM by Google for research assistance and content organisation
- Claude 3.5 Haiku by Anthropic for idea generation and structural refinement
- Grammarly for grammar checking, style enhancement
The final MS Word document represents a collaboration between human authorship and AI assistance, with all AI contributions being reviewed, edited, and integrated through human judgment. Although this declaration aligns with many of the principles of transparency discussed in the work, it is experimental.
Leave a Reply