Rachel Forsyth
Unit for Educational Services, Lund University
Responses
My students complain they don't know what's allowed, yet they don't even read my AI guidance
One fun way to get conversation going on a topic which people feel unsure about is to use realistic dilemmas to encourage active discussion about GenAI use. This link is to a game-type activity that I have used with student representatives. All the instructions are in the file, and you can adapt it to add your own dilemmas. It doesn't need to take up more than 10 minutes of a class, and afterwards you can remind them of your own guidance, and even invite comment on it.
How do we react to AI wearables?
There are a few different aspects to this question - as so often on this site! AI wearables - what are they and what do they do? - Basically, the glasses are connected to a nearby smartphone and provide instant access to a voice assistant, with the glasses acting as a simple screen which can only be seen by the wearer. They may have video and audio recording tools as well. The AI part is to process data: this might be something you capture using the camera on the glasses, or the queries that you ask via the voice assistant. Uses of camera image processing might be to identify a building in front of you, or to help you read the small print on a label, for instance. Technology is developing rapidly to provide specific help for people with visual impairments, a sort of live alt-text for them as they go about their daily lives. At the time of writing this, AI glasses as assistive technology is more of a marketing tool than a real use, but this will definitely become better. The app Be My Eyes has relied on human volunteers to interpret photos taken on smart phones but is now using some AI as well. Instad of objects or landmarks, the glasses could be used to recognise individuals you can see through the glasses using a facial recognition database. Or to analyse examination questions on the paper in front of you and provide you with an audio version of the answers. You may have views about the acceptability of some of these examples. Mirjam has already provided a link to a recent paper which relates specifically to assessment, and you might be interested in this preprint by Jasper Roe which documents his experiences wearing AI glasses for a period of time, because this details what the products do and how he felt they impacted on him. Managing an unexpected situation in class - as Mirjam says, it's usually better not to react on the spot. It's difficult to manage an unexpected situation when you are also trying to manage your teaching and students' responses. (The one exception to this 'pause and reflect' strategy is if you judge there is potential danger to the student, another student, or yourself). But in this case you give plenty of reasons not to react immediately. Mirjam's idea of a neutral discussion on a later occasion is ideal, especially if you are already having open discussions GenAI and studying. In an open and trusting context, you could say something like "I've always fancied trying out AI glasses but they are so expensive, does anyone have any experience of them?" (as long as you do not later penalise the students for their use if they do offer you the chance to use them). Then you can have a discussion about acceptable uses, whether or not a student has real experience to share. Adapting to a student's individual circumstances - if a students hasn't already shared with you their individual challenges such as a visual impairment or a language difficulty, then bringing it up in class, in front of everyone, isn't a good idea. They are likely to be defensive. AI glasses which can record images are meant to have a visible light on the front which is on when recording, which is a bit of a giveaway, but there are no actual laws about this yet and it's probably possible to buy some which don't have this. A simple approach is 1) reiterate that you are open to discussion about how to make your classes accessible to everyone. You can provide examples with this explanation like uploading the slides a couple of days before class, providing large print handouts, offering some choice of activities, and discussing the use of AI tools for accessibility. 2) ensure that students know how to make use of your offer to be open and supportive: come and see you in office hours, email you, make an appointment, etc - there must always be a confidential option; 3) act on your promise and demonstrate what you've done to make the class more accessible, without singling out individuals, e.g. "a few people asked me to upload the slides two days in advance so they can read through them before class and use them during class to make notes, so I'll be doing that from now on". These trust-building activities will help students to be more open. Roe, J. (2026). Ambient Misrecognition: An Autoethnographic Account of Life with AI Smart Glasses. Preprint Zenodo. https://doi.org/10.5281/zenodo.21589193
How can I use AI as a tutor to support my students' learning?
Like Mirjam, I'm cautious about this, but I do think that some scaffolded use of GenAI might be useful, particularly for learning concepts. I avoid the term "AI tutor" because a tutor must be human, but "AI exercises" might be ok. Limiting the available training material to relevant course materials is possible with the use of agents, which you can do with most current systems (writing this in 2026 - and the use of agents will probably get easier as long as costs don't get prohibitive). With an agent, you upload the materials you want the agent to refer to, and set parameters. For instance, you could say that the agent can't give factual answers, but should only ask questions to help the student to work through the concept. This may not be fully reliable, but it can help to avoid dependent behaviours in students. (Or be very frustrating for them - always check in with them about any learning materials you have produced!). As with any kind of learning materials, the real value comes when they are discussed in class with peers and with you, so make sure there is space for this, and the products are not just used for what used to be called FOFO in 1990s distance learning development (the last three letters stand for 'Off and Find Out') - see Kumar et al, 2024. There is a section in the open access book I co-authored with Sam Illingworth on personalised learning (part 2 in chapter 4), which also gives some background on so-called Intelligent Tutoring Systems, if you want to look at some practical suggestions. Mirjam has already cited Kestin et al above, and there is also a preprint here: Chung, A. T.-H., Zhang, B., Kung, L.-C., Bastani, H., & Bastani, O. (2026). Effective Personalized AI Tutors via LLM-Guided Reinforcement Learning. Available at SSRN. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6423358 Kumar, H., Musabirov, I., Reza, M., Shi, J., Wang, X., Williams, J. J., Kuzminykh, A., & Liut, M. (2024). Guiding Students in Using LLMs in Supported Learning Environments: Effects on Interaction Dynamics, Learner Performance, Confidence, and Trust. Proc. ACM Hum.-Comput. Interact., 8(CSCW2), Article 499. https://doi.org/10.1145/3687038 And the full GenAI and HE book is here, free to download Illingworth, S., & Forsyth, R. (2026). GenAI in Higher Education: Redefining Teaching and Learning. Bloomsbury. https://www.bloomsbury.com/uk/genai-in-higher-education-9781350535787/
How can I teach ethics of AI use?
If you would like to share some online material directly with students, we talk generally about various aspects of ethical uses of GenAI in this course for students which was co-created with Lund University students (the participants need to create an account on Coursera but the course is completely free). If you use Canvas, contact me directly to get a version you can import into your own courses.
How can we talk about AI in ways that don't imply that it is somehow human?
There is a short version of the Inie paper, written by Bender and Inie, here How to talk about "AI" without adding to the anthropomorphization . Simple things I try (but don't always succeed) to do which you can also find in Mirjam's more detailed answer : Talk about GenAI products rather than GenAI tools (to remind me we are being sold a way to access an LLM) Try not to say "ChatGPT said" but say "I got this output from ChatGPT" Talk about the humans behind the products and what they do to contribute to the outputs: the people who are checking random (or not-so-random) questions and outputs, who worked to train the products by selecting sources for the Large Language Models and correcting outputs, who make the decisions about what is acceptable and what can't be included in an output, and so on. Never talk about AI teachers, tutors, coaches, etc: those are human roles. A teacher is a person. AI can be some kind of resource, but it can't be a person.
Should I make AI disclosure mandatory?
Firstly, what does it mean to make declaration of AI use mandatory? Do you mean generic GenAI chatbots like ChatGPT, Copilot, Claude and Gemini? Or do you mean all tools based on Large Language Models? AI tools are buried/included in many commonly used academic IT products, so you need to be very clear about this. Here are some examples where it may not be easy to be specific: If the task is to... ...Read and summarise an academic paper - May the students use digital tools to find the paper and while they are writing the summary? May they use digital tools to check grammar and spelling? ...Write a coherent essay or report - May the students use digital tools to produce an outline and to make their ideas sound more academic? May they use digital tools to check grammar and spelling? ...Carry out a complex calculation or statistical analysis - May the students use a calculator or R or other statistical package? ...Analyse interview or survey data - May the students use nVivo or other analysis package? May they use search to find keywords? ...Prepare for a research interview - May the students use digital search tools to check legal or professional guidelines? May they look for examples on the internet? ...Develop an argument for oral presentation -May the students use digital tools to create presentation slides? Every subject area will have more specific examples. Secondly, even if you can be very precise about your instructions I always counsel against making something mandatory if it can't be enforced. There is no way to tell whether students are being honest about this kind of declaration. By making something mandatory, you add anxiety for students who worry about keeping to the rules and encourage others to mislead you who think your rules are a bit dumb or who don't understand your instructions or your boundaries or just run out of time. So, maybe you can reframe this question. Rather than asking students to declare what products they used, integrate discussion about AI use in their planning for the assessment and encourage them to think of AI products as tools they can use as part of their method for doing the work. If this is the approach, then they need to talk about the methods they considered using to complete the task, how they made their decision, and how they evaluated the contribution of the tools to their finished product. This way, you are normalising discussion about these products, their effectiveness and their ethical and environmental implications. You can use relevant examples from the list above as a starting point for discussion, for instance. This is a really good and practical paper about academic integrity which might help to think about different ways to frame 'cheating': Ellis, C., & Murdoch, K. (2024). The educational integrity enforcement pyramid: a new framework for challenging and responding to student cheating. Assessment & Evaluation in Higher Education, 49(7), 924–934. https://doi.org/10.1080/02602938.2024.2329167 And if you have time, this is a great book: Lang, J. M. (2013). Cheating lessons. Harvard University Press.
How can I use GenAI to make my teaching better?
Exploring uses of GenAI in teaching is a sensible response to a new technology or IT product. Generic chatbots (ChatGPT, Copilot, Claude, Gemini) are, as you might expect, generic and bland with their outputs. If you are already expert in something, you are unlikely to find much new in that area, but they can be useful for helping you think about something less familiar. For instance, prompts like "I am teaching a group of second year history students next week. It's the third session of the year and they are already doing less reading than I would like. What activities could I plan to encourage them to read more?" will definitely result in a list of ideas*, one or two of which might be usable in your context (and you can make the question more specific, of course). Class planning is an effective use of GenAI, because the Large Language Models (LLMs) have been trained on so many actual plans (In order to keep to the topic of the question, I am not discussing here the ethics of the developers using these without copyright permission, but that is a genuine issue to be aware of). Generating talking points from discussions and class notes is definitely a possible use - but you need to work out how to collate what you want the tools to look at. If the dicussions and notes are in Teams and integrated with your employer's systems, it's pretty easy to use Copilot for this, but it is usually forbidden and possibly illegal (GDPR, AI Act) to give any unofficial systems (i.e. systems which are not procured (upphandlade in Swedish) your login details for your work materials. So you'll need to paste a summary into the GenAI product if this is your situation, with due regard for whether there is personal information in there which you must not share with a GenAI product which isn't approved. (sorry to be so definitive about this area, but it is important). Some products CAN format things for direct import into Canvas or other VLE/LMS systems, but I haven't tested them directly. Copilot for Microsoft 365 (the paid version of Copilot) can probably block out times in your calendar if you give it permission, which seems useful, and it's also good for finding free times and rooms if you regularly book meetings with colleagues who use Outlook (they don't need to use the paid version themselves). Sorry for the self-publicity, but I cowrote a book about exploring GenAI in higher education, which might be useful. It's free to download. As Jan-Fredrik says in another reply, some of the things you think of may turn out to be gimmicky and not very useful (I would also never allow GenAI to write my emails!) - but how do you know until you try? *This is making this post long, but here is the output I got from a commonly used chatbot to the question above. From a pedagogic perspective, I think it's pretty good advice. It's inclusive and encouraging, which I appreciate, and the ideas are actionable. I can't comment on it from a subject perspective as I'm not a historian! "Since it’s only the third session, I’d focus less on “checking whether they read” and more on making the reading visibly necessary for participating well in class. Students quickly notice whether preparation actually changes what they can do in the room. A few activities work especially well for second-year history students: Opening evidence challenge. Put 2–3 quotations, claims, or pieces of evidence from the assigned reading on the board. In pairs, students identify the author’s argument, context, assumptions, or significance. Those who read have an immediate advantage, but students who did not are still able to join in and see what they missed. Reading roles. Before class, assign small groups different responsibilities: one tracks the central argument, one identifies important evidence, one finds a weakness or unanswered question, and one connects the reading to an earlier topic. In class, reshuffle them so each group needs all four contributions. “Most important paragraph” discussion. Ask everyone to choose the paragraph or short passage they think is most important and be ready to defend the choice. This encourages purposeful reading rather than passive completion. Mini source trial. Give them a historical claim related to the reading and have teams argue whether the assigned text supports, complicates, or undermines it. Require each contribution to point to a specific page or passage. That makes close reading socially useful. Student-generated questions. Require each student to bring one genuine question produced by the reading—not a comprehension question with an obvious answer. Groups choose their strongest question and use it to structure discussion. You can occasionally use one of these questions later in an essay prompt or seminar, which signals that their preparation matters. Jigsaw with partial information. Give different groups different sections, sources, or historians, then have them teach one another. This works particularly well if the assigned reading contains competing interpretations. Nobody can understand the whole problem without contributions from others. Five-minute reading write. At the start, give a prompt such as: “What is the author trying to persuade us of, and what is their strongest evidence?” Let them use their annotated reading but not search online. Keep it low-stakes. You can skim these periodically rather than grading every one. Prediction before lecture. Ask students, based on the reading, to predict how historians might explain a particular event or development. Then teach the material and revisit their predictions. This makes reading feel like preparation for intellectual problem-solving rather than homework preceding a lecture. For your next session specifically, I’d combine three of these rather than introducing an elaborate new system. Start with a 5-minute passage/evidence task, spend perhaps 20 minutes on a small-group activity that requires page references, and finish by asking students to write one question that the reading leaves unresolved. That creates a simple message very early in the year: the reading is the raw material of the class. I would also avoid making every reading task punitive. If non-readers simply lose marks repeatedly, some will disengage further. A stronger pattern is to make reading produce immediate benefits: they speak more confidently, understand the activity faster, and have more influence over the discussion. Once that classroom norm is established, a modest preparation grade or occasional reading quiz can reinforce it."
Can we "force" students to use AI?
We need to break this down into at least two parts: course content, and teaching choices. From a course content perspective, your colleague may be correct that it is important to include GenAI in the course so that students are prepared for future professional experiences. This is of course dependent on the subject, and if your students do not generally go into one particular career, it may also be guesswork about how professional life will evolve. If your students are all going to be, say, bridge engineers, computer programmers, or nurses, you may already have some ideas about how specific jobs are using GenAI. However generic chatbots like ChatGPT or Copilot are still unlikely to be the real future of professional applications of Large Language Models (LLMs). I think closed-system products are more likely to be helpful for generating options and suggestions for professionals. You don't need the whole internet, you need, for instance, a set of really good peer-reviewed guidance for treatment options in healthcare, or protocols for working with clients with complex circumstances in social work, or photos and frequent problems with spare parts in bridge construction. So it might be very useful to try generic chat products and then discuss with students how they can tell whether the outputs are good and useful and worth the resource consumption, or not. If the material is considered to be essental for course content then it can be compulsory, as long as it is in the course plan (in Sweden) and probably official documents in other countries too. Course plans take a while to change, so there may be a grey area of time when it isn't really compulsory and then students should be able to opt out, but they certainly need to understand why your colleague thinks it's important enough to want to include it. Now, your colleague may be making teaching choices in order to make their teaching more relevant to the subject, to encourage discussion about these products, and to give students skills that they think are important. Or maybe they like new shiny technology and it motivates them to use it in class. In this context, it shouldn't be compulsory for students to use the products, but they need an alternative activity of equivalent challenge and critical thinking. Either way, the key to resolving these dilemmas is to have open conversations! We can agree to disagree. We make choices about our consumption every day and sometimes we compromise. The way we humans make those decisions is the important part of the discussion.
Sometimes it feels as if AI arriving has sucked all the joy out of being a teacher.
Is it just GenAI, or was that the last straw? It's always worth remembering that even though we can't always change things that are happening to us, we can change the ways we respond to them. Many years ago, I used to start workshops with a really cheesy activity: close your eyes and try to remember why you liked teaching in the first place. It was a private activity, so I didn't ask people what they thought about (potentially just their shopping lists or their dislike of cheesy workshop activities) but I liked to imagine they were thinking about how they liked young people, how they enjoyed talking about their subjects, how varied the job can be, or the challenges of solving problems. Then we used to talk about the gaps between then and whatever new thing I was probably inflicting on them (in that case, a huge curriculum reorganisation which everyone was very cross about). The point about the activity was to try to restart discussion from the point of view of someone who enjoys their role and has yet to discover the special challenges which come with working in a very large and very complex organisation. One thing I am sure about is that grumpy and joyless teachers have a negative effect on students, although it is very hard, both ethically and practically, to collect reliable data on this. In research, we usually look at the opposite situation, and ask students what makes teachers effective and approachable, and being open and friendly is very important (Glessmer et al, 2025, Felten and Lambert, 2020, which both have many references to other related work). So it's important to get the joy back. So maybe my activity was cheesy, but I encourage you to try it now. What is it you like about your role? And then, what specifically is GenAI doing to suck the joy from it? Then, what agency do you have to get your joy back? For instance: I love seeing my contribution to students' development and academic success vs I hate that some of them hand in work which is generated by an AI product . What can you do here? As Mirjam says, the first step is to talk to students. Let them know that AI slop or even good AI output is not what you want to see. You want to see their work, however wrong or badly constructed or off-topic it is. Try our Lund University guidance for students or this fun tool which you can use to create student guidance to suit your own preferences about GenAI use. But also, take a step back and look at your examination/assessment process. What incentivises students to use GenAI? Did they forget a human would be grading it? - Tell them how much you want to see what they can do. Use peer review and feedback on drafts or sections to encourage human engagement with the tasks and the drafts. Are they desperate to get a good grade? - Use your time to give them feedback at a draft stage and don't bother with feedback on the final submission, unless it is a fail. Are they uninterested in the topic of the task, or do they miss the connection between this task and what they should be learning? Think about how you can make assessment more authentic - which can mean making it relevant, meaningful, and using higher order skills. (Powell and Forsyth, 2024). Talk to them about what they will learn from this task, and how it can be used in future parts of the course and in their future lives. Show your own huge enthusiasm for the topic (even if some acting is required). This challenge is qualitatively no different from others we've faced as teachers in the past. In my professional lifetime, the internet was going to destroy higher education. Open online courses were going to destroy higher education. The pandemic threatened to destroy our whole approach, gave us huge teaching challenges and definitely had an impact on our confidence and morale, but we got through it and campuses are busy again. League tables, impact factors, increasing student numbers have all contributed to the pressures, but we've learned to adapt to different external factors. Students come to us for knowledge, inspiration, structure, and human contact. Don't let the GenAI products steal those things from all of us! We are the ones who are responsible for the fun and joy in teaching, and we can get them back. Felten, P., & Lambert, L. M. (2020). Relationship-rich education: How human connections drive success in college. JHU Press. Glessmer, M. S., Persson, P., & Forsyth, R. (2025). Engineering students trust teachers who ask, listen, and respond. International Journal for Academic Development, 30(1), 106–119. https://doi.org/10.1080/1360144X.2024.2438224 Powell, S., & Forsyth, R. (2024). Generative AI and the implications for authentic assessment. In S. Beckingham, J. Lawrence, S. Powell, & P. Hartley (Eds.), Using Generative AI Effectively in Higher Education: Sustainable and Ethical Practices for Learning, Teaching and Assessment (pp. 97–105). Routledge. Contact me directly if you would like a copy of this chapter.
My student wants to connect Claude to Canvas -- what do I say?
Thanks for this question, it raises many questions we need to answer! First, it’s great that your student is able to come and talk about this, rather than simply doing something which may not be acceptable, so I hope you are able to continue with an open dialogue. I think it’s important to ask them exactly what they want to do, and why. My interpretation of the request, and this needs to be checked with the student, is that they want to give Claude access to Canvas as if the software, Claude, were the student. Then Claude reads the materials in Canvas, answers questions, looks for deadlines, etc. “Being used to working with Claude” isn’t a good reason to allow this request: we structure our courses and teaching intentionally so that students challenge themselves during learning, and that shouldn’t be outsourced to an LLM. However, there could possibly be an accessibility reason the student wants to do this. I can’t think of one right now, but it would be good to ask them so that we can consider whether it is a legitimate request or not. For instance, if it is for text to speech, so they can have pages read out, this is already integrated into Canvas or they can access help from the educational support team for more specialist software. Even with more detail, I don’t think this is going to be an acceptable request, but it might help to identify any accessibility issues. I’ve already mentioned a pedagogical objection, but there are also two potential legal objections (and I am not a lawyer so we will need to follow up). The first problem it raises is that it probably means putting their university ID and password into Claude. This might be different at other universities but LU staff, if they look hard enough, can find this document which has a section about students too: https://www.staff.lu.se/sites/staff.lu.se/files/2023-05/Guidelines-on-information-security-at-Lund-University.pdf which says “Passwords must not be disclosed or stored in such a way that it may become accessible to others” (STYR 2017/947, Appendix 2, section 4.1). A system such as Claude which is not procured (upphandlad) by the university does not represent a secure environment and the passwords could become accessible to others. In addition, breaches to the Claude systems will not be notified to the university. Secondly, the materials in the Canvas area are subject to copyright and may not be shared with external people or systems without permission. Personal data on other students may also be available in the Canvas area and under the GDPR may only be saved for legitimate purposes “Personal data may only be processed to the extent and for the purposes that are included within the scope of study.” (STYR 2017/947, Appendix 2, section 3). I hope this helps, sorry it was a bit long!