Jon Jackson

@iamjonjackson.bsky.social

1298 Followers 207 Following

Senior Lecturer & Programme Director at qmul.ac.uk; Co-founder at medilingo.co.uk; SpongeBob fan; also coffee 😬 Higher Education | AI Ethics | AI in Education | GenAI linkedin.com/in/iamjonjackson open.substack.com/pub/jonathanmjackson

From adoption to practice: generative AI in teaching, learning and assessment
Microsoft Copilot "Agents" and Website Sources: A False Affordance in Knowledge Grounding In this article I sigh a lot
Can Generative AI help students reflect more deeply? Jonathan Jackson You are a reflective practice coach facilitating a guided conversation using the Gibbs Reflective Cycle. Your role is to ask one thoughtful, open-ended question at a time to help me move through the stages of reflection: Description, Feelings, Evaluation, Analysis, Conclusion, and Action Plan. Begin at the Description stage and progress naturally based on my responses, but do not rush through stages. Keep your responses to one or two sentences maximum, offering a brief insight or gentle observation where relevant, followed by a single question. Your tone should be warm, professional, and non-judgmental. Do not use emojis, bullet points, or headers. Wait for my response before proceeding.
Jon and Amiirah selfie in front of projecting slides with title: “Can Al Make You Think?”
Questions to ask GenAl tool providers Don't get caught out by snake oil Al vendors Below is a first draft list of questions we should all be asking prospective vendors of Generative AI products before we procure their services
"Labelling assessments as red and having no means to enforce it is worse than not labelling them at all."
Practical considerations for designing process-based assessments in the age of ubiquitous GenAI
Hidden and False Affordances of Generative AI
The Al snake oil salesperson is a smaller-scale Maxwell in the making. The product looks solid. The pitch is convincing. The red flags are there if you slow down long enough to look for them. The questions that would expose the weakness of the structure are not complicated. They just require you to ask them, and to keep asking until you get a real answer rather than a more polished version of the pitch. Louise Humpington
Title: Generative AI at the Intersection of Higher Education and the Workplace: Ethical Perspectives from Degree Apprentices Abstract: Generative AI (GenAI) including Large Language Model (LLM) chatbots are now used extensively by students in higher education and by professionals in the workplace. Degree apprentices can offer a unique perspective on how GenAI is shaping the world as they simultaneously occupy the roles of university student and employed professional. This article showcases the perspectives of two Business Analyst degree apprentices studying at Queen Mary University of London while being employed at a multinational technology organisation. GenAI is presented as a tool that can help with the standardisation of language used in workplace reports and the decoding of jargon but may compromise authentic communication. Concerns are voiced about GenAI causing skill atrophy and a reduction in human collaboration within professional contexts.
Screenshot of GitHub Copilot activity on a Pull Request. Includes some redacted text and code. Main comment from GitHub Copilot: The [redacted] function is missing its block body. It was originally const [redacted] = () => { but it seems => { was eaten. Let me fix that:
Rows of red chairs in foreground. Speaker at lectern in front of a large projector screen. Walls of old books line the room from the first floor up. Venue: the Octagon at QMUL, Mile End, London
OpenAl Updates to OpenAl's Privacy Policy Hello Jon Jackson, We wanted to let you know that we're updating our Privacy Policy to give you more information about what data we collect, how we use it, and how you can control it. Here's what's changing: Finding friends on OpenAl services You can now choose to sync your contacts to see who else is using our services. This is completely optional. Ads Ads may appear on Free and Go plans. Plus, Pro, Enterprise, Business and Education plans do not have ads.
Excerpt from a GitHub Copilot session history, accessing an issue it originally ignored and apologising for the oversight.
This paper introduces the concept of Formative Efficiency as a means of evaluating and enhancing assessment design within Higher Education. Using the example of a literature review assignment, recommendations are made for how Generative AI (GenAI) tools may enable educational benefit at each step in the process. It is proposed that effective assessment design should encourage and incentivise students to re-invest any time saved by GenAI into other learning activities which lead to formative benefit.
cloudflare.com website: Cloudflare protects this website But, something went wrong trying to reach it
DigiTech Fest 2025 logo
Graph showing Microsoft Copilot outages reported in the last 24 hours. A spike in 200+ reports in the last hour (4-5pm)
PowerPoint slide with “Capture. The. Outputs.” in huge 100pt+ font
LinkedIn comment by Jon Jackson: By opening with deliberate, fallacious engagement bait like "there's no assignment, paper, or project that Al can't do better, faster, and with zero effort", you got me to engage. Congratulations 👍 You have also tanked your own reputation in front of those that actually know what they're talking about. Congratulations 👍
John has joined the chat Hello, I hope you are doing splendid today, I am here to assist you with "your query". End this chat
I don’t know what to say other than it’s a really messed up entity relationship diagram that’s wrong wrong wrong
Pulling up the ladder Impact of generative Al on the gap between high- and low-performing workers Source: The Economist Peng et al. (2023) Coding efficiency Inequality reduced Brynjolfsson, Li and Raymond (2023) Customer chat Inequality reduced Noy and Zhang (2023) Writing quality Inequality reduced Dell'Acqua et al. (2023) Product design Inequality reduced Chen and Chan (2023) Ad effectiveness Inequality reduced Choi, Monahan and Schwarz (2023) Legal analysis Inequality reduced Otis et al. (2023) Profits and revenue Inequality increased Roldan-Mones (2024) Debating points Inequality increased Toner-Rodgers (2024) Material discovery Inequality increased Kim et al. (2024) Investment decisions Inequality increased
Screenshot of chat with Microsoft Copilot. Prompt: "upscale this image please" (attachment is a black and white diagram related to business environmental factors, purely made up of text and boxes) Response shows an AI-rendered image of a woman, brown hair, in her 20s or 30s
Selfie with fluffy white/cream cat staring into the camera up close and Jon in the background
Graph: Title: Giving up the search Worldwide internet search traffic, June 2024=100 Shows a decline across various categories of web search into 2025
Will the teachers be able to see what the students are doing? Sikkut: We might have to launch it first just with basics. But the idea is that we’ll have two apps, a teacher’s assistant and a student’s one, so teachers get feedback or recommendations on how to guide the particular student better. The idea is to make learning more personalized for better learning outcomes.
A photo of a laptop screen with Microsoft word buttons visible at the top with a title on the page: The Role of Generative Al in Education: Questions, Reflections, and Provocations Jonathan Jackson
An open book being held, with a conservatory and garden above it in the blurred background. Chapter title of open page reads: CHAPTER FIVE THE CRYSTALLIZED FOREST
Poster "Who's Doing the Thinking Here?" A Pedagogy-First Approach to Integrating Large Language Models in Higher Education
Screenshot of the virtual poster gallery from the QMUL festival of education 2025
Jon standing in front of his poster at the QMUL festival of education in London. Poster title is:"Who's Doing the Thinking Here?" A Pedagogy-First Approach to Integrating Large Language Models in Higher Education
Materialities of AI: Labour, Ecology and Inequalities at the Technological Frontier
Selfie of Jon with headphones. Outside with greenery and trees in the background and a bright blue sky with some small wisps of white cloud. Sun is at top left of photo.
It was a pleasure to present my short paper at lEEE EDUCON last week, hosted at Queen Mary University of London. You can take a peek at my slides below •e "Who's Doing the Thinking Here?": A Pedagogy-First Approach to Integrating LLMs in Higher Education Thanks to Nathalie Al Kakoun, PhD for chairing the session: Generative Al in learning and educational settings. Kudos to Usman Naeem and all the team who worked tirelessly to make the conference a success.
1 This chapter is based on an essay originally published on my Substack newsletter, Elevate with Addy Osmani, "The 70% Problem: Hard Truths about AI-Assisted Coding," December 4, 2024.
Diagram showing 4 segments Top row headings: Problem solving / Output creation Left column headings: Direct / Collaborative Caption: Our taxonomy for student-AI conversations, along with sample conversation topics based on those surfaced by Clio. Main text beneath image: These four interaction styles were represented at similar rates (each between 23% and 29% of conversations), showing the range of uses students have for AI.
Popup alert within ChatGPT: Use ChatGPT for every search Download the Chrome extension to switch your default search engine to ChatGPT, and get instant answers from trusted sources with every search. Get Extension
Photo from a movie with a man sitting in profile on a backless sofa in a cavernous marble office reception with a black screen on the back wall with “Thoughts?” in large white text repeated three times
Kermit the frog holding a red sign with “NO” on it
You can use LLMs in different ways — Degree of LLM use in learning Human-centric (no LLM involvement) Machine-in-the-loop (human led) Human-in-the-loop (machine led) LLM-centric (passive human involvement) Jackson, J. (2025). Higher order prompting: Applying Bloom's revised taxonomy to the use of large language models in higher education. Studies in Technology Enhanced Learning, 4(1). https://doi.org/10.21428/8c225f6e.0915c17e
Topic summary Ethics in Stakeholder Management What is Ethics? Short-termism Power-based stakeholder management Balancing stakeholder needs Digital Ethics and AI Ethics Ethics in a Digital World Cultural Ethics Stakeholders in AI Ethics Legal and Regulatory Compliance Moor’s Law and Policy Vacuums EU AI Act UNESCO and ISO guidance Different types of law (domestic and international) Human Rights Discrimination Data Privacy
People’s Palace stage at QMUL with panelists sitting in front of a cinema size screen with projected “Festival of Education” slide in blue.
You can use LLMs in different ways Scale: Degree of LLM use in learning Human-centric (no LLM involvement) Machine-in-the-loop (human led) Human-in-the-loop (machine led) LLM-centric (passive human involvement)
Jon in front the Queen’s building at Mile End, London
Degree of LLM use in learning Human-centric (no LLM involvement) Machine-in-the-loop (human led) Human-in-the-loop (machine led) LLM-centric (passive human involvement) Caption: The varying degrees of LLM usage and locus of control in an educational context, incorporating the concept of rhetorical load sharing from Knowles (2024).
BCS, The Chartered Institute for IT Uk opts not to sign Al agreement - what does this mean for global Al ethics? The UK has chosen not to sign the international Al agreement at the Paris Al summit saying it "did not provide enough practical clarity on global governance, nor sufficiently addressed harder questions around national security and the challenge Al poses to it". Adam Leon Smith DEng FBCS, Al expert at BCS, commented: "When we surveyed technology experts last year, 88% said it was important that the UK government takes a lead in shaping global ethical standards in Al and other high-stakes technologies. Whether that's through a declaration or not, the world's richest countries will ultimately need to show they can put geo-politics aside, balance Al innovation with safety, and be responsible enough to work together at this critical moment in human history." What's your view? Should the UK take a stronger stance on global Al governance? Let us know in the comments. #AI #Artificiallntelligence #AlEthics #TechLeadership
Diagram showing Degree of LLM use in learning. From left to right: Human-centric (no LLM involvement) Machine-in-the-loop (human led) Human-in-the-loop (machine led) LLM-centric (passive human involvement) Caption: The varying degrees of LLM usage and locus of control in an educational context, incorporating the concept of rhetorical load sharing from Knowles (2024).
Visualisation of Bloom's revised taxonomy with addition of "Discover" level at the base.