Generative AI Use Codes

Generative AI Use Codes

By Stephen Thomas, Enhanced Digital Learning Initiative (EDLI)

The following is a proposed system of “Generative AI Use Codes” (GAUC) for academic assignments to provide clearer communication between instructors and students. These can be used to communicate the allowed level of generative AI assistance and desired degree of citation in academic tasks. The codes are meant to be simple and easy to use, reminiscent of the approach of Creative Commons licenses. There are two parts to the code: Part 1 communicates the role of AI in the task, and Part 2 communicates the desired attribution of the work requested.

Part 1: Generative AI Use Codes (GAUC)

GAUC-0: No Generative AI Allowed

  • Symbol: AI 🚫
  • Description: Students are not permitted to use generative AI in any capacity for the assignment. 

GAUC-1: Generative AI for Brainstorming Only

  • Symbol: AI ⛈️
  • Description: Students can use generative AI for brainstorming ideas, but the final content must be entirely their own. 

GAUC-2: Generative AI as a Reference

  • Symbol: AI 📚
  • Description: Students can use generative AI as a reference, similar to how one might use a textbook. However, direct output from the AI should not be included verbatim in the final assignment. 

GAUC-3: Generative AI for Editing and Refinement

  • Symbol: AI ✍️
  • Description: Students can draft their own work and use generative AI tools to edit, refine, and polish their content. The initial ideas and content must originate from the student. 

GAUC-4: Collaborative Creation with Generative AI

  • Symbol: AI 🤝
  • Description: Students can collaborate with generative AI to create content. While students should be actively involved in the creation process, they can interweave their own content with content generated by the AI. 

GAUC-5: Unrestricted Generative AI Use

  • Symbol: AI 🌍
  • Description: Students can use generative AI in any capacity, including generating the entirety of the assignment with the AI. They’re encouraged to experiment and innovate using the technology.

Part 2: Generative AI Attribution Codes (GAAC)

N: No Attribution Required

  • Symbol: 🆓
  • Description: Students are not required to provide any citation or acknowledgment for using generative AI, irrespective of the extent of AI’s contribution.

S: Source Attribution Required

  • Symbol: 🔗
  • Description: Students are required to mention the AI tool or platform they used (e.g., OpenAI’s GPT-4), but no specific citation format is mandated.

C: Comprehensive Attribution Required

  • Symbol: 📝
  • Description: Students should provide a comprehensive citation, detailing not just the AI platform/tool, but also specifying parameters, prompts, or any other specifics of how the AI was utilized.

R: Reflection on AI Use

  • Symbol: 💭
  • Description: Beyond merely citing the tool, students need to include a short reflection or description of how the AI was used, its influence on the outcome, and any human-AI collaborative dynamics involved.

Implementation:

Example: On assignment sheets or syllabi, faculty can employ both the GAUC and GAAC codes side by side, for instance, “GAUC-3-C” or “AI✍️📝”. This would indicate that students can use generative AI for editing and refinement, and they need to provide comprehensive attribution for the AI used.

Educational Materials: In addition to the code, it would be beneficial to provide students with a brief guide or overview of the GAUC system, explaining each code and its implications. This could include examples of how to cite or reflect on AI use appropriately.

Honor Code Integration: The concept of proper attribution, even to AI tools, should be ingrained in academic integrity guidelines. Stressing the importance of honest and transparent communication regarding AI assistance aligns with principles of academic honesty.

Faculty Discretion: While these codes provide a structured approach, faculty should retain the discretion to make specific clarifications or exceptions based on the nature of the assignment or the objectives of the exercise.

GAUC – 4S – OpenAI. (2023). ChatGPT (Aug 3rd version) [Large language model]. https://chat.openai.com/chat

Stephen Thomas

Dr. Stephen Thomas is a faculty member and the Associate Director for the Center for Integrative Studies in General Science at Michigan State University and the Digital Curriculum Coordinator for the College of Natural Science. He provides expertise for the EDLI team in pedagogy, curricular reform, and visual thinking.