Session 2 for 2026 from ATAIN with a presentation from Professor George Siemens on 'How AI changes practices in higher education'.
Abstract of the presentation:
After several decades of bold proclamations and
unending hype of the future of learning, from Web 2.0 to open courses to
flipped classrooms, artificial intelligence arrives in our university
classrooms at a moment of change fatigue across the sector. This presentation
will explore how AI impacts learning and knowledge practices, focusing on which
specific tasks, now done by humans, are most amenable to AI. System level
Implications of the core change of “what is done by human cognition and what is
done by artificial cognition” will be explored.
Notes:
Mark Nichols introduced George.
Began with the conundrum represented by the arrival of AI. Too much is changing and things are moving very fast. Almost impossible to anticipate what is going to happen. AI said to replace software developing but the data does not hold up on job replacements in this area.
We are currently moving from chatbots to agentic AI. Many platforms being launched and many people using, experimenting, evaluating. Check podcast - Software developers provide the idea but AI does the coding. Examples out there where coding is being done by AI by startups etc.
Provided an example of how Claude can work if provided a folder of sources - pdfs, reports, etc. The ppt created is accurate. Therefore AI becoming a work bot. Then put the ppt back into Claude to tidy it up and 'approve' the various items that will be used to build the slides. Therefore, one 'manages' AI .
Started the main presentation with assumptions that we are familiar with Gen AI, multimodal and world models, reasoning models, advances in models and encroachment into human capabilities. The uncertainty of labour market impact and that negative aspects of AI (bias, hallucination, IP abuse, environmental and human harm etc.)
Postulated that with distance education in 2002 - open education helped scale content. Then in 2008 open courses scale instruction and now, perhaps ability to scale engagement.
Benefits - personalised learning, positive influence on learning, reduced planning and admin, greater insight into student understanding but negatives in ethics, need for learning design for AI.
All the AIs are dropping LMMs - Claude for education; chatGPT study mode etc.
However, students tend to rely on AI, rather than learn from or using AI :(
Explained how AI LLMs may work. We can not just send students to Claude/ChatGPT to learn; the value add for education is the structure, curriculum, pathways, learning support etc. which are required to help novices learn specialised knowledge. Therefore connect learner profile and content knowledge, through systems prompt/context window (LLM) and compute personalised curriculum in the form of learning content, learning activities which matches the preferences and profiles of the learner.
Agents in higher education can support:
recruitment/registration; wellness support; guide and direct through university experience; create content; teach and coach; assess and evaluate; research. Therefore agents will be LMS, textbook, assessments, teaching, tutoring, apps etc.
Software engineering provide examples of what will happen. In the last 18 months, companies have been built in a weekend; product speed is accelerating and this will happen across many 'knowledge work'. 'Its building coalitions that work - issue by issue, with partners who share enough ---
Stages of engagement:
personal - join open spaces and contribute (your blog, huggingface, share your learning);
coordinate: across institutions, share governance, share strategies and resources.
systems level:
Agency is important as individuals. We can create with AI but there is importance in first defining, thinking, planning and develop the skills to monitor and track what is available.
Encouraged getting lined up with key AI tools and the infrastructure that supports them. Git, Obsidian and file systems, skills and similar repeatable processes and start taking all the free courses (OAI, Deeplearning.ai, Anthropic, Gemini).
Q & A followed.
As a summary, things are moving rapidly and there is no turning back. As individuals, need to keep up with the play. As institutions, need to be pro-active, collaborate and resource innovation/change.
No comments:
Post a Comment