Course Grading and Generative Artificial Intelligence Usage Policy

This document outlines the suggested grading policy and guidelines for responsible AI use in EMI (English as Medium Instruction) courses. These policies aim to foster accuracy, critical thinking, and language proficiency while helping students understand how to use AI tools constructively.

1. Grading Policy

The grading policy in EMI courses is designed to encourage thoughtful engagement with course material, accuracy in responses, and clear communication. Each component of the policy reflects these objectives and includes specific incentives for quality and diligence.

1.1 Accuracy and Reasoning

Students are expected to demonstrate both correct answers and logical reasoning. Grading will reflect:

Penalties for Incorrect Answers: Errors in responses will incur point deductions. Answers unsupported by course content or relevant theories may receive zero points in specific sections to encourage fact-checking and critical evaluation.
Partial Credit for Reasoning: Partial credit is awarded for well-documented logical steps, even if the final answer is incorrect, as logical explanations reflect understanding and critical thinking

Example: For problem-solving assignments, students may earn half to two-thirds of the credit for clearly explained steps toward a solution, even if the final answer is incorrect. Full credit is awarded only when both the conclusion and reasoning are accurate and logically sound.

1.2 Documentation and Presentation

Clear documentation and organized presentation are essential for effective communication, and generative AI tools make these elements more accessible. The following guidelines apply:

Formatting Standards: Additional points may be awarded for assignments formatted in LaTeX or Markdown, enhancing readability and grading efficiency.
Code Documentation: For coding assignments, students are expected to include comments and explanations for their code, which supports both grading and peer review.

Example: Submissions in LaTeX or Markdown with proper formatting and documentation may receive a 5% grade bonus for readability and professionalism.

1.3 Language Proficiency and Grammar

Clear and grammatically correct English is crucial in EMI courses. Points may be awarded for proficient use of English.

Grammar Incentives: Assignments with minimal grammatical errors and clear communication will achieve full grades.
Academic English Incentives: Additional points may be awarded if students’ writing meets standards of formal academic English.

Example: Submissions that demonstrate proficient academic English may receive a letter-grade bonus on that task.

2. AI Usage Policy

AI tools, such as ChatGPT, can be valuable learning aids when used responsibly. This policy provides guidelines on when and how to use AI to complement original work while maintaining academic integrity.

2.1 Encouragement of Fact-Checking

AI-generated responses are not always accurate. Students are encouraged to use AI critically and verify its outputs. Fact-checking is essential for both accuracy and credibility.

Fact-Checking Credit: If AI assistance is used, students should identify inaccuracies and document corrections. Partial credit will be awarded for showing diligence in verifying information.

Example: For assignments involving research, students should verify any AI-provided references or claims. Points will be awarded for identifying and correcting AI errors, as this reflects critical engagement.

2.2 English-First Interactions

AI tools such as ChatGPT generally perform better in English than in Mandarin. Using English with AI not only improves performance but also strengthens language skills.

Example: Demonstrate a comparison of responses from English and Mandarin prompts in a TA session. Emphasize that English prompts often yield more robust answers to encourage students to follow this guidance.

2.3 Transparency and Originality in AI-Assisted Work

Students must disclose any use of AI in their work, ensuring transparency and alignment with academic integrity standards.

Permitted Uses: Students may use AI for proofreading, drafting code, or formatting assistance.
Transparency Statement: Students must include a brief statement explaining how AI was used in completing each assignment. Misrepresentation of AI assistance may lead to grade deductions.

Example: At the end of each assignment, students should append a brief statement detailing AI usage, such as for drafting, formatting, or idea generation. Failure to provide this transparency may impact grades. Prompts should be archived and made available for further review if requested by the instructor.

3. Common Concerns and Solutions

This section addresses potential issues with AI usage and grading, and provides preventive measures.

3.1 Fake References

To prevent the submission of unverified sources:

Random Source Checks: Spot-checks may be performed to ensure that references are credible and relevant. Fabricated or unverified AI-generated references will result in substantial point deductions.
Invisible Instructions (Less Recommended): To prevent students from blindly copying AI-generated answers, consider embedding instructions in white text within the question prompts. If students fail to verify the prompt and responses, these instructions may reveal inaccuracies upon submission.

3.2 Vague or Loose Answers

AI often produces general responses, especially for complex subjects. AI-generated answers may avoid abstract mathematical statements and lack depth in argumentation. Students should avoid vague answers by:

Specificity in Responses: Students are encouraged to refine AI-generated answers to be specific and detailed, particularly for mathematical or argument-based responses.

4. Initial Policy Introduction

To ensure students understand these policies:

Policy Presentation: Host a TA session or lecture to cover grading and AI policies, providing examples and addressing questions.
Written Documentation: This document, along with a summary email, will provide clear guidance on responsible AI usage and grading criteria.

Acknowledgement

This document was generated by ChatGPT based on my presentation slides and notes for a session at National Taiwan University of Science and Technology (NTUST).