David Hopkins’ Post

🚨 Is Accuracy a Problem When Using Generative AI in Higher Education? Indeed, not all output from GenAI tools like ChatGPT is accurate. However, despite potential inaccuracies, these tools can still be incredibly useful. I think it helps to break things down a bit: 🔹 1. AI for Improving (Student) Writing: For non-native speakers and those refining grammar and sentence structure, AI tools are highly effective. In these cases, accuracy isn’t a significant concern as AI excels at language rules. Since it’s improving clarity rather than generating new content, there’s little room for “hallucinations.” 👉 Inaccuracy isn’t a concern here. 🔹 2. AI for Generating Ideas and Structures: When using AI to brainstorm or develop a framework for a topic, accuracy takes a back seat. AI offers valuable starting points that, while not always complete, are rarely factually incorrect. 👉 Inaccuracy isn’t a concern here either. 🔹 3. AI for Data Analysis: AI can write code and tools like ChatGPT can run analyses using built-in Python capabilities. If the input is clear, AI typically performs well—similar to running the process manually, though you don't have to write all code yourself. While mistakes can happen, with proper oversight, AI can be a powerful aid in analysis, whether you run it on your machine or not. 👉 Inaccuracy isn’t a major concern here, though caution is needed. 🔹 4. AI for Providing Precise Knowledge: This is where accuracy becomes a significant issue. When tasked with providing expert-level knowledge, citations (e.g. for a literature review), or any kind of specific information, AI still struggles much at the moment. While newer models (OpenAIs o1-preview!) and tools like Perplexity are providing citations, they still often lack depth, making them unreliable for tasks requiring precision. 👉 Inaccuracy and hallucinations are major problems here. 💡 Takeaway: AI can be an incredibly valuable tool in higher education. For tasks like writing improvement and idea generation, accuracy isn’t a big issue. But when it comes to detailed research or empirical data, caution is key. ⚠️ What’s your take on the dangers of inaccurate GenAI output?  👇 #AI #HigherEd #EdTech #GenerativeAI #AIinEducation #Teaching #Learning

These AI tools can help you find academic papers, debunk conspiracies, and more. #edtech #AI

How ChatGPT works? ChatGPT is a generative AI product that uses large language model (LLM). Training ChatGPT is similar to how babies learn language, involving three main phases: unsupervised learning, supervised learning, and reinforcement learning. First, in the pretraining phase using unsupervised learning, ChatGPT learns like babies do when they listen to others to pick up language patterns. The model is given a large amount of internet data to understand language. It learns words, sentences, and context by processing this data. This helps the model develop a foundational understanding of language. Next, in the fine-tuning phase using supervised learning, ChatGPT learns through direct teaching, similar to how babies learn when adults correct them. In this phase, the model is trained on a dataset of human-labeled prompts and responses. This helps the model produce more relevant and accurate responses by learning from the labels provided. Finally, in the reinforcement learning phase using human feedback, known as Reinforcement Learning from Human Feedback (RLHF), ChatGPT learns through rewards, like how babies receive claps, smiles, or treats when they respond appropriately. The model's responses are evaluated by humans who rate their quality. Based on this feedback, the model adjusts its responses to optimize for maximum rewards, producing more human-preferred and ethical answers. #generativeai #chatgpt #artificalintelligence #machinelearning #deeplearning

When you require students to disclose the role of #AI as a routine part of an assignment, you also open up the avenue for #students to realize that the tool may not actually have helped them. In our culture, we’ve become so accustomed to viewing failure as a bad thing that #younglearners avoid taking risks. But requiring open disclosure sends the message that it’s OK for them to try something new, and not succeed at it. Learning requires friction. Marc Watkins shares how to get students to disclose and evaluate their own usage of tools like #ChatGPT on The Chronicle of Higher Education: https://ow.ly/Pcnu50TCmKb

Generative AI tools like ChatGPT can revolutionize education, offering every student a smart and infinitely patient research assistant. Here's how they can utilize it to best effect (and you can too)... #ai #genai #chatgpt #copilot #education

“Education will always be people-centered. But generative AI offers a unique opportunity to rethink the role that technology can play in how we teach, how we learn, and how we think about the very nature of education.” Read more insights from Elon University - Martha and Spencer Love School of Business educator Haya Ajjan, about how to successfully harness the power of AI to engage students and personalize their education: https://lnkd.in/ewMgjjhK #AI #AIinEducation #ChatGPT #GenAI #HigherEducation

A lot of hand-wringing about #ChatGPT in teaching circles, with good reason. But: “I prefer, as a publisher and educator, to focus on the unique contribution human intelligence will continue to make in our collective space for the foreseeable future and focus on how #AI will support and enhance the uniquely #human.” – David Parker Read: https://lnkd.in/enKYBYdu

Talk, talk, talk, talk....and with AI in the picture, we can add Chat(GPT), Chat(GPT), Chat(GPT)....an endless babble of noise. Yes, the opportunities of what AI can do for teaching and learning are many, but the cacophony of noise is overwhelming. Where are the pedagogies and the simple approaches for using this? In today's article, we share a simple, 4-Step Framework to Prepare Learners for a ChatGPT World. Check out this great article by Oguz A. Acar on his PAIR framework. Have you tried something like this? https://lnkd.in/g-s7UQzw? #ai #artificialintelligence #chatgpt #pedagogy #edtech #education #highereducation #aiineducation #copilot #k12 #assessment #bard #gemini #LearningAndDevelopment #TeachingTips #EducationLeadership #OnlineLearning #EdLeaders #InstructionalDesign #EdChat

Great tips on assignment design when using AI. #highered #ai #genai #chatgpt #studentlearning #studentsuccess #assessmentforlearning #assessment

At Tilburg.ai, we are continuously exploring how to responsibly integrate AI into education. A common question we receive is: How do you get started with this? To support students and professionals, I’ve written an article titled ‘How to Start Using ChatGPT: A University Student’s Guide.’ This article offers an introduction for anyone looking to start using ChatGPT in their studies or work. We also offer additional resources, including: - A course on how to ask questions, prompts, using ChatGPT. - An FAQ addressing common questions about AI. Feel free to explore the article and resources below to learn more about how AI can support your studies or professional development.