EduExe’s Post

After a recent talk on leveraging GenAI in higher education that I gave, two familiar questions came up: 1) How should we handle essays that show signs of being generated by GenAI? 2) What do we do when students know more about GenAI than the faculty? The first issue is tricky; the second perhaps not so much. First off, as far as I know there’s currently no tool that can reliably identify if an essay was written by GenAI. Accusing students without sound proof can lead to legal problems (more on this, see links below). Instead, what we can do is discuss expectations upfront. If a text seems like a copy-paste from ChatGPT, it’s likely because it’s written poorly or the style feels off—which is both bad style and bad GenAI usage. If used well, GenAI should enhance the style, not make it worse. If we can't legally flag GenAI usage reliably, it's probably better to avoid the dilemma altogether: either permit and support(!) GenAI (with the expectation that good usage should lead to excellent results both with regard to accuracy and style!) or design assessments that avoid its use, like closed-book exams or oral formats. As for students knowing more about GenAI than their instructors—I’m absolutely assuming they do! And I think it’s great. Importantly, though, there’s often a big disparity in skills within a classroom. We can use this to create a level playing field where students help each other learn, and where faculty can engage with students on more equal footing, as both faculty and students have unique knowledge to share. This kind of dialogue can bridge faculty concerns—like accuracy, citation practices, and biases—with practical insights from proficient users! ➡️ If your institution or department is having these debates and you'd find some external input helpful, please let me know. 🤝 Some resources that might be useful On inability to detect GenAI generated output reliably: https://lnkd.in/e62KS25R (slide 12 and onwards) https://lnkd.in/epSg9SEb For an intuition how GenAI output is created: https://lnkd.in/ep5djyZn https://lnkd.in/e3BsEueZ #GenAI #HigherEducation #TeachingWithAI #AIinEducation #GenerativeAI #EdTech #AcademicIntegrity #StudentLearning #FutureOfEducation #ArtificialIntelligence #FacultyDevelopment #DigitalLiteracy #LearningTogether #AIUsageInClassroom

I saw a trending post on Threads where a professor shared that: ✰A student showed him an AI recorder ✰The student admitted it was so he "doesn't have to worry about paying attention to the lectures" ✰The prof was impressed by the outcome of the AI ✰The prof reflects that in a few years it will just be computers recording and summarizing what he says (presumably instead of students attending) ✰There will be a point when AI just does the homework, too. I don't know anything about the professor who posted, and this is not a jab at him in any way. But aside from the final point being what every teacher has been screaming about for the last 20 months, the post reminded me of a few important things: 1️⃣ Many of us here on LinkedIn or other EdTech friendly spaces are in a bubble. Out there in the real world, a lot of teachers are still just trying to teach and haven't really considered the impacts of AI thoroughly. 2️⃣ Traditional lectures are DEAD. Some teachers just don't know it yet. 3️⃣ A lot of profs don't have actual backgrounds in pedagogy. This was a shock to me when I joined Higher Ed. There's a whole swath of teachers that got jobs because they're subject-matter experts - but subject-matter expertise does not a good teacher make. 4️⃣ We all need to spend a LOT more time making sure students understand the VALUE of learning and critical thinking for their own sakes. Too many feel like it's just a series of check-boxes they have to achieve in order to get a job. If you're out there talking, giving workshops, helping your colleagues, or whatever else... KEEP GOING. It's going to take a long time for people to see what's happening, so just be careful that you don't let the bubble of tech-enthusiastic teachers here influence your impression of where most people are. There's a long way to go, and people need your help! 💪💪 #AIinEducation #TeacherDevelopment #Pedagogy #Teaching #ArtificialIntelligence

#ConversationalAI is changing how students interact with their universities, offering instant, natural support for every question. From academic advice to administrative assistance, this technology ensures every student feels heard and supported 24/7. Ready to see how? 👇 Explore the impact of ConversationalAI in #Higher Education (details in the comments below). #Blog #DRUIDAI

An important article by John Warner, who's one of the best minds out there when it comes to discussions about writing in an educational context. It's on the importance of limits, and the virtues of being a "hardliner" when it comes to incorporating/giving into the use of LLMs in educational spaces. At the heart of the matter is ethos. Universities should be spaces of criticism and reflection. That's one of the fundamental things that make them different from corporations, which in some contexts may benefit from "moving fast and breaking things." But if universities eagerly dive head into the latest fad, they carry the risk of breaking some pretty important stuff. Some of the hardlines that Warner recommends we do not cross? * LLMs should not be used as writing tutors or graders. After all, they can neither read nor write (this is an objective claim). * LLMs should not be used for peer-review (peer being the operative word). * LLMs should not be used to replicate human data in experimental studies that make claims about human behavior (this shouldn't even be a point of debate). * LLMs should not be mobilized to replace the educational labor force--mainly faculty and graduate students. Proclamations about "mission" and "values" come cheap. So cheap that you can ask ChatGPT to generate them. Meanwhile, as Warner warns, crossing the lines outlined above is compelling evidence that a university has lost sight of its institutional mission as it obsessed over its institutional operation. It reveals an institution that prioritizes volume over quality, scalability over interaction, simulations over genuine experiences. Is that what we should aspire toward when it comes to research and teaching? Is the goal to transform our educational spaces into something akin to Facebook? Might as well turn off the lights if that's the case... https://lnkd.in/dmr46tV2

correcting students' midterm exams First observation: their use of language resembles that of politicians! very vague and general which is not the correct and complete answer to a question but can be 'interpreted' to respond to the question! Second observation: No, they are not politicians, they didn't study and they have low English skills. Third impression: Nope, generative AI is drastically (even compared to last year's exams) reducing their capacity to write and argue properly. Fourth fearful thought: are we going towards a global illiteracy hence an easier manipulation of the masses??! Note: The above does not apply to a few well-versed exam papers, though. #education #universities #highereducation

I’m torn when reading this 😨 If this helps those more fearful of #AI 🤖to adopt new tech, then this could be beneficial 🚀 If I were asked a year ago to see how AI could help feedback on and test young people’s public speaking skills - my answer would have been “kind of” - today my answer is absolutely. 👀 What the sector needs is a vision of HOW AI can help the sector and EEF could lead this 💭Imagine if every young person had a virtual tutor with the knowledge of the learner, access to the worlds information on every subject, the rubric for assessment and a user friendly way of interacting and providing feedback and learning experiences. AI will sit behind this. Perhaps the most valuable outcome of the study will be more people comfortable with AI 🤖

📚 𝗨𝗖𝗟𝗔 𝗶𝗻𝘁𝗿𝗼𝗱𝘂𝗰𝗲𝘀 𝗔𝗜-𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗲𝗱 𝗰𝗼𝘂𝗿𝘀𝗲 𝗺𝗮𝘁𝗲𝗿𝗶𝗮𝗹𝘀   UCLA is breaking new ground in winter 2025 with a comparative literature course featuring materials developed by AI. Covering literature from the Middle Ages to the 17th century, this course is developed by using #Kudu, a textbook platform created by UCLA professor Alexander Kusenko.   𝗖𝗿𝗲𝗮𝘁𝗶𝗻𝗴 𝘁𝗵𝗲 𝗰𝗼𝘂𝗿𝘀𝗲 𝘄𝗶𝘁𝗵 𝗞𝘂𝗱𝘂 Professor Zrinka Stahuljak provided Kudu with her lecture notes, PowerPoint slides, and YouTube videos from previous iterations of the course. Kudu then transformed this content into a complete package: textbooks, homework assignments, and TA tools. Creating the course takes about 3–4 months, but Kudu helps to streamline the process, requiring professors to dedicate no more than 20 hours to tasks like reviewing and editing materials.   𝗔𝗜 𝘄𝗶𝘁𝗵 𝗴𝘂𝗮𝗿𝗱𝗿𝗮𝗶𝗹𝘀 Unlike public tools like ChatGPT, Kudu operates in a controlled, closed-loop environment. It only responds to questions based on course content provided by the professor, ensuring academic integrity and reducing the potential for misuse. For #students, it offers a safe platform to ask questions anonymously—questions they might otherwise hesitate to ask in class—and provides answers drawn directly from the professor’s input.   𝗙𝗶𝗿𝘀𝘁 𝗼𝗳 𝗮 𝗸𝗶𝗻𝗱 Originally developed for UCLA science courses, Kudu’s expansion into the humanities demonstrates the platform’s versatility. Stahuljak’s course represents Kudu’s debut in UCLA’s Humanities Division, though the platform is already in use this term in a social sciences course   Looking ahead, Stahuljak plans to use Kudu in additional courses, noting that the platform allows her to spend more time guiding students through primary texts and teaching essential skills such as critical thinking and analysis. Teaching assistants (TAs) will also have more capacity to help students improve their writing—a crucial yet often challenging element to address in large classes.   #AIInnovation #ResponsibleAI #EthicalAI #EthicAI # #AIResearch #TechForGood #TrustworthyAI #AIinEducation #FutureOfLearning #UCLA #education #universities   (🔗 to article in comments)

Yesterday, we delivered our first workshop on Artificial Intelligence in Education at Baku-Oxford School, and one major takeaway was addressing teachers’ anxiety about the rapid pace of technological advancements. Many teachers feel overwhelmed by the sheer number of challenges AI presents, but expecting to tackle them all at once only adds to the stress. Instead, we suggested a more manageable and constructive approach. I introduced an activity adapted from Duncan Foord's excellent book, 'The Developing Teacher' (highly recommended!). The activity, called “Coffee Stains,” focuses on identifying and addressing small but persistent irritations in the workplace. Here’s how we applied it to AI-related concerns: 1️⃣ Teachers identified three “coffee stains”—their top three AI-related concerns. 2️⃣ For each, they considered: ✔️ What can I do about it? ✔️ How and when can I address it? 3️⃣ Finally, they added these actions to their to-do lists to progressively check in on their progress. This method encourages teachers to prioritize, think critically, and actively seek solutions, whether it’s reaching out to colleagues for support or exploring training opportunities. The goal is not to solve everything at once but to focus on manageable steps that reduce anxiety and promote confidence in dealing with AI. Find out more about Duncan Foord’s 'The Developing Teacher' at https://lnkd.in/ecc4yVKP #ArtificialIntelligence #AIinEducation #TeacherDevelopment #ProfessionalDevelopment #InnovativeTeaching #TeacherTraining #EducationTransformation #TeacherSupport #DuncanFoord

While AI excels in many areas, critical thinking and personal connection remain uniquely human strengths. By embracing experiential and reflective-based pedagogical methods, educators should empower students to engage deeply with their studies. This approach fosters critical thinking skills and encourages students to draw connections to their personal experiences, enriching their learning journey. AI serves as a valuable aid for students, provided they grasp the ethical considerations, seek guidance from instructors, and approach tasks with genuine intellectual curiosity. Let's ensure that AI complements rather than replaces human intellect, nurturing a generation of mindful learners prepared to navigate the complexities of the future. #AIinEducation #CriticalThinking #EthicalAI

📝 Introducing Pregrade Providing students with meaningful essay feedback can be time-consuming. That’s where Pregrade steps in! Our AI tool grades essays, drafts personalized feedback, and even corrects spelling and grammar, while teachers and educators alike maintain control by reviewing and adjusting the AI’s evaluations. Think of Pregrade as your AI grading assistant. Just like the wealthiest schools hire human grading assistants, you can now give the AI a stack of essays and a rubric. In a few minutes, Pregrade will return detailed feedback, saving you up to 80% of your grading time. No more waiting weeks for results—Pregrade delivers efficient, equitable feedback without the complaints or delays. Find out more here: https://pregrade.ai/ #EdTech #Education #AIAssistance #AI #Pregrade #Productivity #Grading #Teacher #AIforEducation Joon Nak Choi, Kylie C., Ivy Shi, Aritro Mukherji, Ian Reed, Robert Manjo, Lawrence Lai, Mark Cohen