Generative AI in teaching & learning

AI in Higher Education: Pedagogy and practice

This page contains my developing notes on the use of AI/Chat GPT in my own practice. Broadly speaking there are opportunities and potential threats from the way students can engage with such tools and the same probably goes for a wide range of administrative tasks in Higher Education practice.

Video explainers for students: Appropriate use of generative AI in higher education

A short guide (3 min) and a full version (27min) designed for students to watch and in support of their engagement with AI tools such as ChatGPT but with guidance on how to avoid submitting coursework that is deemed to have made inappropriate use of such AI tools.

Slideset used above: I made a shareable copy of the slides I used above and it should be downloadable at this link.

An introduction to the use of AI in HE

This whole page is an introduction to the use of AI but in this section are included some of the key terms and a few introductory resources.

A Brief Glossary

Artificial Intelligence (AI) involves using computers to do things that traditionally require human intelligence. AI can process large amounts of data in ways that humans cannot. The goal for AI is to be able to do things like recognize patterns, make decisions, and judge like humans. AI is not new and AI is a much broader field than LLMs and ChatGPT. It can be used in many ways AND we have all been using it for years (e.g. music streaming, search engines, social media, online shopping, online advertising)
Large language models (LLMs) are more recent advances in AI that work on human languages. It understands and generates text in a human-like fashion. It's an advanced chatbot that can also trawl the internet and quickly produce succinct answers to complex questions.
ChatGPT (Google Bard, Bing Chat etc) are all apps/tools that allow users to engage with a LLM and ask it questions. These are the user interfaces.

Further Introductory Resources

What is ChatGPT, the AI software taking the internet by storm? - an 8 min clip BBC News in Dec 2022 https://youtu.be/BWCCPy7Rg-s
JISC (2023, May) A Generative AI Primer https://nationalcentreforai.jiscinvolve.org/wp/2023/05/11/generative-ai-primer/
Sabzalieva, E., & Valentini, A. (2023). ChatGPT and artificial intelligence in higher education: quick start guide. UNESCO https://www.iesalc.unesco.org/wp-content/uploads/2023/04/ChatGPT-and-Artificial-Intelligence-in-higher-education-Quick-Start-guide_EN_FINAL.pdf
Lingard, L. (2023). Writing with ChatGPT: An Illustration of its Capacity, Limitations & Implications for Academic Writers. Perspectives on Medical Education, 12(1), 261. https://doi.org/10.5334%2Fpme.1072

AI tools and apps

This section includes a list of tools and apps that allow relatively easy use of AI.

AI/Large Language Models

ChatGPT - https://openai.com/blog/chatgpt - v 3 is FREE (inconsistency access, not as clever), v4.0 is ~$20/month (and is always accessible and more powerful)
Google Bard - https://bard.google.com/ - use your google log-in, you might need to register to be part of it. Very good access, similar to ChatGPT v3.5
Bing - You must be using Microsoft Edge as your browser. Look for the "chat" tab on the Bing homepage. If you have trouble accessing Chat...make sure you are signed in with your Microsoft log-in. If you can't use it due to SafeSearch, go into settings and switch away from Safesearch. Other tips include...make sure that you have the latest version of the Bing app installed, clear your browser’s cache and cookies, and restart your device, if the issue persists, try uninstalling and reinstalling the app, consider setting Microsoft Edge as your default browser.
Perplexity - https://www.perplexity.ai/ - Free to use, although a paid option allows more access to the CoPilot feature. This tool appears to be accessing more up-to-date searches of the internet than GPT-3 and makes use of GPT-4 within the free access, albeit in a limited manner. It also appears to offer direct links to its sources in its output.
- A comparison of ChatGPT, Bard and Bing - https://www.theverge.com/2023/3/24/23653377/ai-chatbots-comparison-bard-bing-chatgpt-gpt-4

Apps that use AI

Consensus - "Ask a question, get conclusions from research papers" - https://consensus.app/search/ (Seems good for signposting to good sources AND providing snippets from each. )
Elicit - Another tool that can be asked questions and will provide references and signposting to papers https://elicit.org/ [A review of Elicit - May 2023]
SciSpace - https://typeset.io/ - ask a question and get summaries from key papers. Add columns of information as required. [A review of SciSpace - Aug 2023]
Wonders - https://app.readwonders.com/ - another app for finding information
Evidence Hunt - https://evidencehunt.com/ - Ask a question and it will be turned into a literature search of PubMed. Outputs are listed as abstracts that are presented with a PICO analysis overlayed so each key element of the abstract is highlighted a different colour.
The Literature - https://www.the-literature.com - similar to Evidence Hunt in that it searches PubMed but results can be presented in a range of ways like SciSpace
Research Rabbit - find new papers and make connections - https://www.researchrabbit.ai/
Connected papers - explore connected papers in visual graphs - https://www.connectedpapers.com/
Carrot2 - Similar to above - https://search.carrot2.org/#/search/web
Scholarcy - An AI powered article summariser, builds interactive flashcards of articles - https://www.scholarcy.com/
Paper Digest - An AI powered article summariser - https://www.paper-digest.com/
Laterl.io - AI article summariser - https://www.lateral.io/
Inciteful.xyz - Find papers, make connections, export to Zotero - https://inciteful.xyz/
Quillbot - A free paraphrasing website - https://quillbot.com/
Ryte.me - an AI writing tool - https://rytr.me/
Paperpal - an AI writing tool - https://paperpal.com/
writefull - an AI writing tool - https://www.writefull.com/
Vizcom - produce a line drawing, refine the image description with words and allow the app to render a full version of your desired image - https://app.vizcom.ai
Pi.ai - An AI chatbot that you can talk to. On iOS it can hear your voice and maintain a conversation, On Windows and Android, it responds to typed chat but can respond with voice. Maybe use it to get feedback on your initial ideas https://pi.ai/talk

Academic Malpractice and AI

ChatGPT and other Large Language Models (LLMs) can be used to create original written content that students may use in their assessments.

The output created by these tools is potentially coherent enough for it not to be detected by academic staff members, or traditional text-matching software used to detect plagiarism BUT the use of these tools may not necessarily be considered plagiarism if students use them for specific aspects of their studies and possibly are transparent in how they have been used in any submission. There could however be potential for a breach of academic integrity policies and the line between acceptable use and a transgression is one that needs exploration and discussion.

Key Points

AI is new in the practice of academic staff, and students - there are many ways in which AI can be of great use to students and staff alike and it presents great opportunities, just as the development of the internet itself did many years ago. Tools that are widely available are evolving fast and staff and students that are not engaged with such tools will not understand their potential or pitfalls as well as those that are engaged with them.
ChatGPT/Bing/Bard present a potential threat to the originality of a student's submitted assessments: While these tools present opportunities to help students, they are also tempting to use as shortcuts, or worse, substitutes for independent thinking. There is, however, a line between acceptable use and inappropriate use but this will take exploration, reflection and discussion to identify. Students will need to be exposed to such conversations, just as they are around plagiarism and originality issues already.
The ability of Turnitin to identify AI malpractice is "evolving". Much work is being done to find a mechanism of AI writing detection however concerns with false positives and the transparency of the identification tool have led to doubts about its efficacy. The May 2023 update from Turnitin is worth a read on this topic. (Summary: there are appropriate concerns, don't rely solely on the tool but continue to prioritise professional judgement and talk to students in question).
Many early examples of AI misuse have been identified and dealt with in existing plagiarism process and policy - poor referencing, limited attribution of sources, inconsistent writing style, repetition of key text are all examples of potential AI misuse, and familiar to HE staff as well as within existing guidance on poor scholarship.
Talking is key - talk to other staff, and when fears are raised talk to the students in question and ask about how they completed the work, and explain your fears.

Signs a student’s essay might have been written using AI

Adapted from: https://medium.com/@marginaliant/5-signs-a-students-essay-has-been-written-using-ai-a569235ab1be

Lack of Clarity and Coherence - One of the most common signs that an essay has been written using AI is a lack of clarity and coherence. AI-generated essays often lack a clear thesis statement and tend to be disjointed, with no clear connection between the different paragraphs.
Repetitive Content - Another sign of an AI-generated essay is repetitive content. Because AI algorithms rely on pre-existing data to generate essays, they often end up repeating the same points over and over again, without adding any new insights or ideas.
Grammatical Errors - Despite the advances made in AI technology, grammar and syntax remain a challenge for most AI algorithms. As such, essays generated using AI are often rife with grammatical errors and awkward sentence structures.
Lack of Originality - AI-generated essays are notorious for their lack of originality. Since they rely on pre-existing data, they tend to regurgitate the same ideas and arguments that have been made in the past, without adding anything new or innovative to the conversation.
Unnatural Language - Finally, essays generated using AI often feature unnatural language that sounds stilted and robotic. This is because AI algorithms are still not advanced enough to replicate the nuances and complexities of human language, resulting in essays that lack a human touch.
Incorrect Referencing - Early versions of ChatGPT/LLMs have struggled to provide accurate citations. They can look correct, but may well be reconstituted and adapted citations that use real authors and journal titles but meld them together into manufactured sources. This inability will no doubt be addressed over time however, examination of citations and references could demonstrate poor scholarship and potentially use of AI.

Considering AI in module assessments

Watch a seminar

How do you design assessments now that students can use ChatGPT? What does ‘innovative authentic assessment’ even mean? Do I have to re-write all my modules? This online webinar looked to unearth the true meaning of authentic assessment and give attendees real life examples of assessment design that account for ChatGPT but still offer an accurate way to assess what a student has learned. Speakers included: Jan McArthur, Lancaster University; Matthew Glanville, The International Baccalaureate; Dr Thomas Lancaster, Imperial College London and Chair of QAA’s Academic Integrity Advisory Group; and George Bryant-Aird, Edge Hill University. The video below is hosted alongside other related content at https://www.qaa.ac.uk/membership/membership-areas-of-work/academic-integrity/chatgpt-and-artificial-intelligence#

Watch the seminar https://youtu.be/leWtOVL1nt4

Ideas for adapting assessments

Use AI to produce an output based on the existing essay title or task. Take that output and hand that to students. Ask them to critique it in class, then ask them to improve it for their submission.
Set a task for students that explicitly requires the use of AI in its completion. Ask students to track how they use it. Then ask students to critique what AI was or was not capable of and submit a commentary on the use of AI in X[field of work

Getting the most out of AI: Tasks, prompts & ideas for teaching and learning

Tasks/Prompts for students

Get help starting your work: Ask it to provide an outline for your work, e.g. an essay plan for...
Get help with referencing: Paste in a bunch of references and ask it to ensure they are all in line with APA7 (for example)
Get help with critical analysis: Ask it to outline a critical analysis of a particular theory, approach or technique (the response will need thinking about, it won't come with specific underpinning references, it might be in the wrong tone, but it might give you some thoughts on key potential critiques).
Get help with the quality of your written prose: Once you have written some text that includes the key points that you want to communicate, and your references you could try asking it to "improve the quality of the writing in this piece of academic writing". The output will need thinking about but there may be some useful changes. it might even tell you what changes it has made which could help you improve your own ability to write in the long term.
Get help consolidating topics and revising/reviewing: Ask it to produce 10 multi choice revision questions on X, or "review questions on X".
Get help with your writing: Ask it to, "Review and if required, improve, the reporting verbs in this paragraph - insert para".
Get help producing an action plan in response to your previous submission: Ask it "What can I do in response to the feedback on this essay - insert feedback".
Raise your game preparing for that dream job: Ask it to write a cover letter for a job and include a link to your CV, a link to the job advert and a link to the JD [Google Bard accepts links to google docs]".
Get help with your stats: give it some data or a link to a spreadsheet and ask it to perform some analysis, or ask it what analysis might be suitable.
Get help with your qualitative research: Ask AI to produce an interview guide for a semi-structured interview in a study that uses qualitative research methods. (this will not be good enough to use but might be a good start that is better than staring at a blank page).

Tasks/Prompts for Tutors

Use it to produce low-stakes tests and review/revision questions: Ask it to produce 10 multi choice revision questions on X, or "review questions on X".
Use AI to produce many varied examples: Students often need many examples when learning complicated concepts but these can take time to produce.
- Copy this text prompt into Google Bard. The reply should ask you some questions about your needs and then you can be more specific about the topic and audience for your examples. "I would like you to act as an example generator for students. When confronted with new and complex concepts, adding many and varied examples helps students better understand those concepts. I would like you to ask what concept I would like examples of, and what level of students I am teaching. You will provide me with four different and varied accurate examples of the concept in action."
Using AI to Provide Multiple Explanations: Teaching involves logical and coherent explanations and effective explanations lay the groundwork for foundational knowledge that helps students build mental maps of topics
- "You generate clear, accurate examples for students of concepts. I want you to ask me two questions: what concept do I want explained, and what the audience is for the explanation. Provide a clear, multiple paragraph explanation of the concept using specific example and give me five analogies I can use to understand the concept in different ways."
Using AI to Assess Student Learning: Several classroom assessment techniques can help tutors and students monitor their learning and understanding of course material. Such assessments can encourage active learning and reflection by asking students to summarize and interrogate their knowledge. These exercises can help tutors but also increase students' engagement and motivation by showing that tutors are responsive to their needs and that their questions and opinions matter.
- You could set up a Google form or doc and ask students any of the following questions... What was the most important idea or concept covered in class today? Why do you think this idea is important? What is the most difficult class concept so far? What did you struggle to understand? What concept or problem would you like to see explored in more detail?
- Then you could feed the responses into an AI tool like Google Bard (you can include a URL from a Google doc and it will read the content). e.g. "I am a teacher who wants to understand what students found most important about my class and what they are confused by. Review these responses and identify common themes and patterns in student responses. Summarize responses and list the 3 key points students found most important about the class and 3 areas of confusion: [Insert material here]"
Using AI to Distribute Practice of Important Ideas: Students need to practice retrieving new information not just once but multiple times during a course. Distributed practice, or having students practice material several times over days and months, is critical to developing robust and flexible knowledge.
- Try a prompt like this: "You are an expert teacher who provides help with the concept of distributed practice. I am teaching the topic of exercise physiology to first year degree students. I have already covered cardiovascular anatomy and respiratory anatomy but I am now teaching the endocrine system. Please provide 4 ideas about how I can include the past topic into my current topic. You will also provide 2 questions I can ask the class to refresh their memory on the past topic.."
Use it to produce short passages of introductory text for VLE pages: The best VLE content contains specific resources like slides or articles to read but also a narrative framework that explains what these resources are and how they fit together. Brief intro text on given topics can help to set the tone of this content and is easily produced by tools such as ChatGPT.
Produce outline meeting agenda and minutes: AI/LLMs can be used to review notes and content such as previous minutes and then produce an outline meeting agenda. They can also produce minutes by inputting notes and links to documents, also potentially using transcripts of meetings.
- [Several of the ideas for tutors were taken from this paper which has further explanations and evidence-based rationale: Mollick, E.R. and Mollick, L. (2023) Using AI to Implement Effective Teaching Strategies in Classrooms: Five Strategies, Including Prompts. SSRN http://dx.doi.org/10.2139/ssrn.4391243

More on prompts and use of AI in academia

Atlas, S. (2023) "ChatGPT for Higher Education and Professional Development: A Guide to Conversational AI." https://digitalcommons.uri.edu/cba_facpubs/548
@hefrteducator (2023) A Teacher's Prompt Guide to ChatGPT aligned with 'What Works Best https://usergeneratededucation.files.wordpress.com/2023/01/a-teachers-prompt-guide-to-chatgpt-aligned-with-what-works-best.pdf
Mollick, E.R. and Mollick, L. (2023) Using AI to Implement Effective Teaching Strategies in Classrooms: Five Strategies, Including Prompts. SSRN http://dx.doi.org/10.2139/ssrn.4391243
Mollick, E. R., & Mollick, L. (2022). New modes of learning enabled by ai chatbots: Three methods and assignments. Available at SSRN. https://dx.doi.org/10.2139/ssrn.4300783
Mollick, E., & Mollick, L. (2023). Assigning AI: Seven Approaches for Students, with Prompts. arXiv preprint arXiv:2306.10052. https://dx.doi.org/10.2139/ssrn.4475995
Compton, M. (live doc) Sandpit: Testing the capabilities of chaptGPT - examples of reading, precising, reformatting text and refs, tabulation, rubric, feedback and marking https://docs.google.com/document/d/1K_UgkLt6--Bqv_FViREBvRzXbD4yR4LzHep15caln4U/edit#heading=h.4uc5kd2g4low
McKnight, L (2022, Oct) Eight ways to engage with AI writers in higher education. THE https://www.timeshighereducation.com/campus/eight-ways-engage-ai-writers-higher-education
Langston, T. (2023, June) Last Night ChatGPT Saved My Life… [A blog about a PG student's use of AI] https://teltales.port.ac.uk/2023/06/26/guest-blogger-tom-langston-last-night-chatgpt-saved-my-life

Ideas for using AI in teaching and learning

Teaching Critical Analysis: Use AI to undertake some classic critical analysis tasks (e.g. "You are a lecturer in higher education. Please provide a critique of Foucault's theory of power from 1998"). Along with other typical critical analysis questions in your area, this would expose students to what we mean by "critical analysis". AI provides the quick answers BUT we have to know the right questions to ask. Another use of AI in the teaching of critical analysis would be to ask it to produce some written work (give it an essay title) and then use a series of critical analysis lenses to reflect on the quality of the work (e.g. quality of sources, descriptive versus analytical writing, use of language such as reporting verbs and hedging words, discussions about the quality of evidence presented, where might the gaps in evidence be etc). This second task provides an opportunity to reflect on what we mean by critical analysis AND exposes the students to the fact that AI is not flawless.

AI and Research

UKRIO (2023, July) "I can’t help falling in love with AI”: chatbots and research integrity. https://ukrio.org/ukrio-resources/ai-in-research/

Additional Reading and Insight for Staff

JISC (2023) AI Case Studies https://nationalcentreforai.jiscinvolve.org/wp/
JISC (2012 & 2023) Artificial Intelligence in Tertiary Education https://beta.jisc.ac.uk/reports/artificial-intelligence-in-tertiary-education
QAA (2023) ChatGPT and Artificial Intelligence https://www.qaa.ac.uk/membership/membership-areas-of-work/academic-integrity/chatgpt-and-artificial-intelligence
Mollick, E. (2023, Apr) The future of education in a world of AI: A positive vision for the transformation to come. https://www.oneusefulthing.org/p/the-future-of-education-in-a-world
- "The first place that AI panic hit was classrooms. AI’s potential for cheating, and for new forms of teaching, is so immediately obvious that I run into people all the time who believe the whole education system is going to collapse. But I actually think the opposite is true: education will be able to adapt to AI far more effectively than other industries, and in ways that will improve both learning and the experience of instructors......"
Mollick , E. (2023, Feb) My class required AI. Here's what I've learned so far. https://www.oneusefulthing.org/p/my-class-required-ai-heres-what-ive
Watkines, R. (2022, Dec) Update Your Course Syllabus for chatGPT https://medium.com/@rwatkins_7167/updating-your-course-syllabus-for-chatgpt-965f4b57b003
Gleason, N. (2022, Dec) ChatGPT and the rise of AI writers: how should higher education respond? https://www.timeshighereducation.com/campus/chatgpt-and-rise-ai-writers-how-should-higher-education-respond
Inside HigherEd (2023, March) A Guide to Generative AI Policy Making https://www.insidehighered.com/views/2023/03/22/ai-policy-advice-administrators-and-faculty-opinion
JISC (2023, May) A Generative AI Primer https://nationalcentreforai.jiscinvolve.org/wp/2023/05/11/generative-ai-primer/
Bath Uni (2023, March) ChatGPT and Artificial Intelligence https://teachinghub.bath.ac.uk/generative-artifical-intelligence/
Bournemouth Uni (2023, Jan) - Artificial Intelligence apps and teaching and learning - https://microsites.bournemouth.ac.uk/flie/2023/01/30/artificial-intelligence-apps-and-teaching-and-learning/
UCL resource for students - Engaging with AI in your education and assessment https://www.ucl.ac.uk/students/exams-and-assessments/assessment-success-guide/engaging-ai-your-education-and-assessment
Edinburgh Napier University QuickGuide - https://staff.napier.ac.uk/services/dlte/Documents/15%20Artificial%20Intelligence,%20Student%20Learning%20and%20Assessment%20Quick%20Guide.pdf
UK Gov (2023, March) Generative artificial intelligence in education. https://www.gov.uk/government/publications/generative-artificial-intelligence-in-education
WonkHE (March, 2023) An avalanche is coming this time. https://wonkhe.com/blogs/an-avalanche-really-is-coming-this-time/
ISTE. Artificial Intelligence in Education: Putting Educators and Students in the Driver's Seat https://www.iste.org/areas-of-focus/AI-in-education
Nerantzi, C., Abegglen, S., Karatsiori, M. and Martínez-Arboleda, A. (Eds.) (2023). 101 Creative ideas to use AI in education. A collection curated by #creativeHE. Graphic Design by Bushra Hashim. CC-BY-NC-SA 4.0. https://doi.org/10.5281/zenodo.8072949
The Russell Group (2023, July) Russell Group principles on the use of generative AI tools in education https://russellgroup.ac.uk/media/6137/rg_ai_principles-final.pdf
SOAS (2023, May) ChatGPT and generative AI – a guide for students https://www.soas.ac.uk/sites/default/files/2023-05/ChatGPT-and-generative-AI-a-guide-for-SOAS-students_0.pdf