AI for learning and teaching
Learning and teaching with AI
AI can contribute to all aspects of learning, teaching and assessment by complementing, supplementing, enhancing, and enriching teaching practices. This can be done via a variety of different AI tools and a diversity of uses and applications, bearing in mind the need for responsible, ethical, and fair AI use for us and for students. Warwick endorses the use of Copilot - for staff and for students.
There are many potential applications for AI in teaching -in class, online, and in synchronous and asynchronous activities providing pre- and post- class learning opportunities. We will cover issues relating to assessment in the next section.
Expand the accordions to engage with some of the key contexts and considerations, discover use cases and find guidance and cautions.
In this section, we consider some of the opportunities presented by Generative AI technologies for learning and teaching. The following examples might be used directly in teaching or to support students to study with AI.
AI can be used to help students develop critical analysis skills, evaluative judgement, and editorial capabilities. Options include:
- Engage students to generate text using AI-generative models and ask them to appraise the quality of the output. The appraisal could look at relevance, authenticity, bias, cohesion, and consistency of the output. Modelling such appraisal process can develop students’ cognitive abilities to evaluate AI-generated output and promote AI literacy. Students could find limitations in the output and develop an understanding of the standards of their written tasks.
- Generate base information using AI tools and have students demonstrate their knowledge by finding, correcting, and explaining errors in the generated information.
- As GenAI outputs become more reliable and sophisticated students can edit AI texts to expand, refine, and improve them. AI models have cut-offs for their training data so asking students to integrate recent news events, research, and/or data.
- Ask students to generate a number of introduction paragraphs, then assess them using a provided rubric or set of criteria. What are the features that make a good introduction?
- Generate a range of example texts that can be used for structured comparison (David Nicol has some useful resources Link opens in a new windowon this). Students analyse the similarities and differences between multiple responses to a question. In medicine, for example, students could compare different treatment plans. In business, they could compare different financial reports. This can lead to insights about the topic being written about, different ways of approaching a written task, and the advantages and disadvantages of AI writing tools.
Text generators can quickly produce scenarios, case studies, personas, dummy data sets, etc to use as example applications of ideas or to use in learning activities such as simulations. They can also create examples of work in different genres, e.g. policy brief, case study, research abstract, and can help students to understand the differences and expectations associated with different assessment tasks.
AI could also be useful for generating a range of examples that are relevant to different settings or cultures (e.g. by specifying a location, nationality, religion, etc.) but these would need to be checked carefully.
Generative AI tools have the potential to be used creatively to advance learning and teaching across disciplines. Examples may include:
- Engaging students to use AI tools to generate texts on a topic for different audience groups and asking them to compare and contrast the language used.
- Using AI tools for creative writing, including in non-creative writing disciplines. For example, students could write the first paragraph, the AI tool could write the second paragraph, and so on.
- Engage AI in a critical debate about a set topic - or extend by asking AI to play the role of a historical figure.
- Use AI to role play e.g. client, stakeholder, patient etc with students directly interacting with AI, or by using to generate scripts to facilitate in-class role play. You could also use AI to develop forum theatre scripts.
- Use AI tools to create choose-your-own-adventure scenarios for students to try out different actions and reflect on potential consequences.
- Using AI tools to support students’ independent study. For example, students could use AI tools to generate responses on study concepts and help them structure their knowledge around each concept by comparing and contrasting. Students should be sure to check the output for accuracy and relevance.
- Using AI tools to implement effective study strategiesLink opens in a new window. Examples include: generating quizzes and flashcards to support retrieval practice Link opens in a new windowand generating prompts to elaborate an idea.
- Using AI tools as a language learning companion Link opens in a new windowfor language learners.
Find more ideas on how AI can be used in teaching Link opens in a new windowfrom the University of Sydney.
Prompt
“Write a postgraduate level multiple choice question about autobiographical memory and ageing. Give four options. Add feedback.”=
Response from Copilot
Question:
How does ageing typically affect autobiographical memory?
Options:
A. Older adults generally show a decline in recalling specific episodic details but maintain the ability to recall general semantic information.
B. Ageing has minimal impact on autobiographical memory; both episodic and semantic memories remain relatively stable.
C. Older adults often improve in recalling specific episodic details due to increased life experience.
D. Ageing primarily affects the recall of semantic information, while episodic memory remains largely unaffected.
Feedback:
- Option A:Correct. Research indicates that older adults often experience a decline in the ability to recall specific episodic details, while their ability to recall general semantic information remains relatively intact.
- Option B:Incorrect. Numerous studies have shown that ageing does impact autobiographical memory, particularly in the recall of episodic details.
- Option C:Incorrect. While life experience can enrich memory, the general trend is a decline in the recall of specific episodic details with age.
- Option D:Incorrect. Ageing primarily affects episodic memory more than semantic memory, contrary to what this option suggests.
Output generated by Copilot 7/02/2025.
Asking Claude to review for accuracy. Suggested that I integrate references to relevant research literature in order to reinforce postgraduate engagement with the literature and generated the following feedback.
Response from Claude
-
- Option A: Correct. Research consistently demonstrates older adults experience reduced episodic specificity in autobiographical recall while maintaining semantic memory (Levine et al., 2002). This pattern aligns with Conway's Self-Memory System model (Conway & Pleydell-Pearce, 2000) and is further supported by neuroimaging studies showing age-related changes in hippocampal function during episodic retrieval (Addis et al., 2011).
- Option B: Incorrect. Longitudinal studies (Piolino et al., 2010) and meta-analyses have documented significant age-related changes in autobiographical memory, particularly affecting episodic recollection. This contradicts the stability hypothesis suggested in this option.
- Option C: Incorrect. While older adults may have richer life experiences, the phenomenon of reduced autobiographical specificity in aging is well-documented. Piolino et al. (2006) demonstrated that even highly significant life events show reduced episodic detail in older adults' recall, a finding consistent with the CAGE (Cognitive Aging and Episodic Memory) framework.
- Option D: Incorrect. This option reverses the established pattern. Research by Nyberg et al. (2012) and others shows that semantic knowledge remains relatively preserved in aging, while episodic memory shows significant decline. This pattern is consistent with the dedifferentiation hypothesis of cognitive aging.
Claude, 07/02/2025.
Checked in with Copilot about the validity and relevance of resources. Copilot added weblinks and suggested additional sources.
Copilot review of references and suggested improvements
Option A: The reference to Levine et al. (2002) is excellent. You might also consider referencing a meta-analysis by Simpson et al. (2023), which provides a comprehensive review of how healthy and neuropathological aging affects autobiographical memory using the Autobiographical Interview
Option B: Piolino et al. (2010) is a solid reference. Additionally, the review by Simpson et al. (2023) also supports the significant age-related changes in autobiographical memory, particularly in episodic recollection
Option C: Piolino et al. (2006) is appropriate. You could also reference the work by Nyberg et al. (2012), which discusses the preservation of semantic memory and the decline in episodic memory with aging.
Option D: Nyberg et al. (2012) is a good choice. The article "Aging and Autobiographical Memory: The Long and Bumpy Road" provides a detailed discussion on the factors affecting autobiographical recall in older adults.
Revised prompt
Write a postgraduate level multiple choice question about autobiographical memory and ageing. Give four options. Add feedback which includes
references to relevant literature.
Response from Copilot
Question:
Which of the following statements best describes the impact of ageing on autobiographical memory?
Options:
A. Older adults typically recall more specific and detailed autobiographical memories compared to younger adults.
B. Ageing primarily affects the emotional content of autobiographical memories, with older adults recalling more negative events.
C. The specificity of autobiographical memories tends to decline with age, while the overall amount of detail remains relatively stable.
D. Older adults show no significant differences in autobiographical memory performance compared to younger adults.
Correct Answer:
C. The specificity of autobiographical memories tends to decline with age, while the overall amount of detail remains relatively stable.
Feedback:
Research indicates that ageing affects autobiographical memory in various ways. One key finding is that older adults often experience a decline in the specificity of their autobiographical memories, meaning they recall fewer specific details about past events. However, the overall amount of detail in their memories can remain relatively stable.1
This suggest that while older adults may struggle with recalling specific events they can still provide a general account of their past experiences. Studies have also shown that the emotional content of memories can be influenced by ageing, with older adults sometimes recalling positive events more vividly than negative ones.2
1. Mair, A., Poirier, M., & Conway, M. A. (2021). Age effects in autobiographical memory depend on the measure. PLOS ONE.
2. Nature Research Intelligence. (n.d.). Autobiographical Memory and Aging. Nature.
Prompt
Create a postgraduate level multiple-choice question about [INSERT TOPIC]. Ensure the question adheres to the following principles of good MCQ design:
- Clarity and Precision: The question should be clear and focused on a specific aspect of autobiographical memory and ageing.
- Plausible Distractors: Include four options, with one correct answer and three plausible distractors based on common misconceptions or related concepts.
- Balanced Length: Ensure all options are of similar length to avoid giving away the correct answer.
- Avoid Absolute Terms: Avoid using absolute terms like "no significant impact" or "always" in the options.
- Positive Framing: Frame the question positively to avoid confusion.
- Vary Correct Answer Position: Ensure the correct answer is not always the same option to avoid predictability.
Additionally, provide detailed feedback for each option, including accurate and relevant references to support the feedback. Use the following structure for feedback:
- Correct Option: Provide a brief explanation and cite relevant research studies or theoretical models.
- Incorrect Options: Explain why each option is incorrect and cite relevant research studies or theoretical models.
Example:
Question: How does ageing typically affect autobiographical memory?
Options: A. Older adults generally show a decline in recalling specific episodic details but maintain the ability to recall general semantic information. B. Ageing has minimal impact on autobiographical memory; both episodic and semantic memories remain relatively stable. C. Older adults often improve in recalling specific episodic details due to increased life experience. D. Ageing primarily affects the recall of semantic information, while episodic memory remains largely unaffected.
Feedback:
Correct option: Research consistently demonstrates that older adults experience reduced episodic specificity in autobiographical recall while maintaining semantic memory (Levine et al., 2002; Simpson et al., 2023). This pattern aligns with Conway's Self-Memory System model (Conway & Pleydell-Pearce, 2000) and is further supported by neuroimaging studies showing age-related changes in hippocampal function during episodic retrieval (Addis et al., 2011).
Incorrect options:
Longitudinal studies (Piolino et al., 2010) and meta-analyses (Simpson et al., 2023) have documented significant age-related changes in autobiographical memory, particularly affecting episodic recollection. This contradicts the stability hypothesis suggested in this option.
While older adults may have richer life experiences, the phenomenon of reduced autobiographical specificity in aging is well-documented. Piolino et al. (2006) demonstrated that even highly significant life events show reduced episodic detail in older adults' recall, a finding consistent with the CAGE (Cognitive Aging and Episodic Memory) framework.
This option reverses the established pattern. Research by Nyberg et al. (2012) and others shows that semantic knowledge remains relatively preserved in aging, while episodic memory shows significant decline. This pattern is consistent with the dedifferentiation hypothesis of cognitive aging.
Here’s an example prompt to try, after which you can have a whole conversation where Copilot will answer what you write (in French or English) in French, followed by an English translation:
Prompt
“Have a conversation with me. Please write your answers in French and then add an English translation. I heard it will rain later.”
Bien sûr, je serais ravi de discuter avec vous ! 🌧️
Sure, I'd be happy to chat with you! 🌧️
Vous avez entendu qu'il va pleuvoir plus tard ? Avez-vous des plans pour la journée malgré la pluie ?
You heard it will rain later? Do you have any plans for the day despite the rain?
Output generated by Copilot 07/02/2025.
Feedback can come from multiple sources, including teachers, classmates, peers, the physical or social environment, and now AI. Carless and BoudLink opens in a new window (2018) champion the need for feedback literacy and developing capacities to appreciate feedback, make judgements, manage affect and take action. Building feedback literacy can support coordination and integration of feedback from a range of intelligences.
Henderson and Philips (2014) challenge us to consider various aspects of inputs:
- the goals of the task (feed-up)
- clarifying what they did well and not so well (feedback)
- what can/should they most productively work on in the future (feed forward)
Consider where in the assessment process feedback occurs. Does it appear incrementally during the production process to guide development or at the end of the process judging the outcome and learning journey while feeding forward to inform subsequent work.
Educator feedback can operate within an ecosystem of inputs from teachers, peers, AI technologies and others. How can we structure feedback rich assessments that work with AI?
Note: It is not permissible to upload student work to an online platform without their consent and without ensuring that the system has appropriate data protections.
Marking represents judgement of quality based on expressed criteria and provides indicators of achievement of learning outcomes.
In evaluating the achievement of learning outcomes, we need to carefully determine what knowledge and skills need to be demonstrated independently by an individual human and which knowledge and skills can be demonstrated collaboratively with multiple intelligences (with classmates or with AI assistance). Similarly, for marking, we need to consider where human judgement is required and where AI can assist. For example, can AI perform spelling, grammar, readability and citation checks while humans perform a check of evidence and argument in a collaborative marking process?
The following are some key considerations as approaches and practices regarding responsible use of AI and marking evolve:
- Privacy rights must be maintained. Do not upload any personal information that does not have adequate privacy protection and contributes data to public training of AI models. For example, do not enter student work into an AI platform without the student’s explicit permission. Even with permission you should de-identify all work to protect privacy.
- Intellectual Property (IP) rights need to be maintained. Students maintain some ownership of their work and we should not upload student work to AI systems that do not have adequate privacy protection and will be contributing the student work to public training of AI models. Copilot would be acceptable due to commercial usage security.
- Accuracy is crucial and we must have confidence that the AI system providing marking advice or marking is generating accurate responses. We need to be able to explain how the AI is making judgements and how we are checking and confirming judgements, maintaining a “human in the loop” for marking with AI processes.
- Transparency is required and we must inform students how AI will be used in marking processes.
- Utility, what parts and aspects of an assessment can AI productively contribute to whether judging criteria or offering feedback?
- Personalisation, how can AI contribute to tailored responses calibrated to our diverse cohorts of students? If AI is used in marking then how can we create human connections that are a backbone of rich university experiences?