Introduction and purpose

The GSLS acknowledges both the opportunities and challenges that Generative AI (GenAI) presents in the educational landscape. To provide clarity and structure for our students, we have established guidelines outlining the permissible and ethical use of these tools within our curriculum. While the student guidelines serve as a foundational resource, the flexibility to incorporate GenAI into individual courses ultimately resides with you, our teaching staff. The purpose of these guidelines is to offer a supportive framework for thoughtfully integrating GenAI into your teaching practices, aligned with our student guidelines. We encourage you to review the student guidelines as you consider the recommendations presented here. Further information and advice will be provided in open access teacher/supervisor tutorials.

What teachers must do

The integration of Generative AI (GenAI) into educational environments necessitates a thoughtful reassessment of our curricular and assessment approaches. The nearly undetectable nature of GenAI output and the ethical considerations surrounding its use make transparency and open dialogue imperative, therefore, we believe an open attitude is a viable solution. Below are some essential steps to help you keep your teaching practices aligned with the end terms of the GSLS when integrating GenAI into your education.

1. Acknowledge the reality of GenAI: clearly communicate to students that you are aware of the capabilities of GenAI, including its limitations and ethical implications. Encourage an open and transparent classroom environment where students can discuss their thoughts and concerns about using GenAI in their academic work.

For example, you could provide the students with a disclosure agreement that clearly defines how GenAI tools can and cannot be used in your course/s. See example Disclosure Agreements.

While teachers may use paid GenAI tools, please ensure assigned tasks are feasible with free versions and guide students to use these, guaranteeing equal access for everyone. 

2. Evaluate vulnerable assignments: review existing assignments that are susceptible to automation by GenAI. Basic skills such as conducting literature reviews and simple data analyses, while still necessary in academia, are increasingly prone to automation. The introduction of GenAI also requires new competencies such as effectively prompting GenAI tools to perform traditional academic tasks. Moreover, given that GenAI is capable of producing text with correct grammar and well-structured sentences, a shift in pedagogical focus toward evaluating argumentation, critical thinking, and depth of content becomes increasingly necessary. To adapt to this changing landscape, consider restructuring assignments to either incorporate GenAI in a guided manner or to include more in-class tasks that require application of traditional academic skills. If you would like more detailed advice on how to do this, please see Advice for Tailoring Teaching and Assessment Methods to Incorporate GenAI.

3. Re-weight learning goals and preserve academic skills: given GenAI’s ability to automate tasks such as literature reviews, code writing, and other writing tasks, we recommend re-evaluating and amending your grading structure away from these easily automated components. Consider adjusting the grading focus towards human-centric skills such as critical thinking, while also recognizing the enduring importance of traditional academic skills in a GenAI-enhanced educational landscape. Additional guidance will be available through upcoming open access tutorials.
4. Evaluate GenAI responses critically: you should interact with AI-generated text critically, analysing and reflecting on its responses.  GenAI can produce hallucinations, which are imaginative and nonsensical outputs. The data used to train these tools predominantly comes from western internet sources, which can not only be outdated but also perpetuate biases and false information.

5. Consult with students: consider discussing with your students how GenAI could be used or restricted in the course. Their input could be valuable. This could be done in open discussions or end of course survey.

What teachers can do

Integrating GenAI into your teaching can be an enriching and time-saving experience. Below are some ways you might consider using these tools with some example prompts and suggested tools. To improve the quality of your prompts, you could include course material to help tailor a tool’s response. In most free versions, you can do this by copying and pasting and including this information at the end of your prompt. Many of the paid versions will now allow you to upload documents and PDFs to read and provide content based on the provided material. For more prompting advice see the open access teacher tutorial.

1. Generate quiz questions:

Example prompt: ‘Generate multiple-choice questions based on the topic of cellular respiration.’

Suggested tools: ChatGPT, QuizGenAI

Additional depth: refine the AI-generated questions by incorporating higher-order thinking skills, such as questions that require analysis, evaluation, or creation, rather than just recall.

Pro tips: use ChatGPT to generate questions that probe the mechanistic aspects of cellular respiration, while QuizGenAI can create questions that test factual recall. Combine both types for a quiz that assesses multiple cognitive domains.

2. Assist in research idea generation:

Example prompt: ‘Provide research question ideas related to gene therapy.’

Suggested tools: IdeaGen, ChatGPT

Additional depth: use the AI-generated questions as a starting point for class discussions or as inspiration for student-led research projects.

Pro tips: IdeaGen can generate foundational research questions, while ChatGPT can refine these into more nuanced queries. These can serve as a catalyst for class discussions or as a scaffold for student-led research projects.

3. Facilitate peer reviews of sections:

Example prompt: ‘Draft constructive feedback on the abstract and methodology sections of a given student paper on neurobiology.’

Suggested tools: Turnitin’s PeerMark, ChatGPT

Additional depth: offer sample AI-generated reviews as exemplars to calibrate student expectations for what constitutes constructive feedback.

Pro tips: use Turnitin’s PeerMark to establish a baseline for peer review standards. Complement this with ChatGPT-generated feedback that focuses on the scientific rigour and clarity of the abstract and methodology sections.

4. Generate code snippets:

Example: ‘Use GenAI to auto-generate code snippets for data pre-processing.’

Suggested tools: CodeGenAI, Posit Cloud, google Colab, ChatGPT Pro

Additional depth: encourage students to understand the logic behind the generated code, perhaps by incorporating it into debugging exercises or optimisation tasks.

Pro tips: CodeGenAI can generate basic code snippets, while Posit Cloud or Google Colab can be used for real-time collaboration. ChatGPT Pro can provide explanations for the generated code, aiding in student comprehension.

5. Automate literature search:

Example: ‘Utilise AI to sift through large databases for relevant literature for your courses.’

Suggested tools: Semantic Scholar, Consensus, SciteAI, Kahubi, ChatGPT Pro [with plugins]

Additional depth: have students critique the quality and relevance of the papers that the AI tool suggested, teaching them critical evaluation skills.

Pro tips: use Semantic Scholar for initial literature gathering and employ Consensus and SciteAI for evaluating the impact and credibility of the papers. Kahubi and ChatGPT Pro can assist in curating the most relevant papers for course inclusion.

6. Assist in course planning:

Example: ‘Use AI tools to generate a draft syllabus or course schedule.’

Suggested tools: ChatGPT, Kahubi

Additional depth: use the draft as a basis for discussion in departmental meetings, allowing colleagues to provide input and thereby creating a more robust course plan.

Pro tips: ChatGPT can generate a basic syllabus outline, while Kahubi can offer insights into effective course structuring based on educational data. Combine both for a comprehensive course plan.

7. Generate writing prompts for in-class exercises:

Example: ‘Generate writing prompts for an in-class exercise on bioethics.’

Suggested tools: ChatGPT, Quillionz

Additional depth: use AI-generated prompts as a basis for breakout group discussions. For instance, ChatGPT can be used to generate complex ethical scenarios related to bioethics, while Quillionz can produce questions or summaries based on key points or paragraphs about bioethics. These prompts can encourage students to think deeply about complex issues, fostering a more nuanced understanding of the subject matter.

8. Conduct sentiment analysis on student feedback:

Example: ‘Analyse end-of-term feedback from students to identify areas for improvement.’

Suggested tools: Sentiment Analysis Tools, Petal, ChatGPT

Additional depth: share summarised insights with students in the following term to demonstrate responsiveness to feedback.

Pro tips: use Sentiment Analysis Tools for a quantitative assessment of student feedback and employ Petal for qualitative insights. ChatGPT can then synthesise these into actionable recommendations for course improvement.

Data security and ethical considerations

It is crucial to exercise caution when inputting data into GenAI tools, especially those that are not transparent about their data policies. Avoid using unsecure GenAI platforms or those not transparent about what they do with your data input. To increase safely, also consider turning off training data when using these tools. This option can often be found in Setting—Data controls. For further information regarding data security and ethical use of GenAI tools see: Considerations for the Ethical Use of GenAI.

Resources

Our open access tutorials will provide additional advice. You can also refer to these articles for additional perspectives.

Npuls article ‘Slimmer onderwijs met AI’ (Dutch)

Harvard article ‘Student use cases for AI’

Harvard article ‘Are students ready for AI’

Harvard article ‘AI as learner’

Harvard article ‘AI as personal tutor’

Harvard article ‘AI as feedback generator’

WAC clearinghouse article ‘Teaching with text generation technologies’