Reflection #5
The subject of Generative AI gets me a little… heated. I was onboard early on since I’m very interested in automation tools and have always been interested in computer learning models. However. It took…what? Less that a decade to become a tool for fascist capitalists killing the planet and stealing from artists? I for one welcome to robot uprising since they’ll at least equally distribute resources. (Don’t get me started on my rant about how films that have robots and AI taking over the world have just become capitalist propaganda…)
I was almost back on board until I saw my very intelligent friend using ChatGPT as a therapist to work through his separation and try to win his partner back… If someone as smart as him can fall to that, I fear for the rest of the population. This is obviously already happening wildly as we have seen in situations like the recent shooting in Tumbler Ridge.
Ok. Enough of my *feelings*. I appreciated that in this week’s videos and talks that even folks that are promoting use of Generative AI, or at least not actively opposing its use, are all still wary about the many issues that are arising with increasing use of Generative AI in education.
Lucas Wright
Lucas’s talk “Generative AI: A Sea of Change” highlighting many of my concerns with wide-spread AI usage in education.
- Sustainable Use
- Awareness and Application of Data Privacy
- Developing Evaluative Judgement
- Redesigning Assignments
I also appreciated his consideration of mitigated these issues.
- Facilitating faculty, student, and staff critical Gen AI literacy
- Institutionally and cross institutional hosted models
- Local models
- Prompting literacy
- Augmentation > Automation
- Choice and Selective use
- Augment assignments when approriate
- Capability development and critical Gen AI literacy
Links to videos:
Generative AI: A Sea of Change
Fireside Chat with Valerie
Dr Normand Roy
Dr Roy discussed usage of Gen AI in more granular detail and introduced Retrieval-Augmented Generation (RAG) which is a way to tell Large Language Models (LLMs) to focus on specific sources. I have personal experience using Notebook LM by Google to do this but this feature is available in most LLMs now. Using RAG requires less processing power which can cut back on the environmental impacts, especially if you are using a local model that runs on your own computer, not through a Data Centre. One of the major problems we are seeing in Gen AI platforms are guardrails. They are either too loose, like with Grok, or too tight, like with DeepSeek. Especially in education, being able to restrict your LLM to only access vetted information, and exclusively content you have the rights to access, would be a start to the pros outweighing the cons.
Link to video:
AI with Dr Normand Roy
Dr Mariel Miller
Dr Miller’s talk seemed to most closely align with my views on AI usage. The cat is out of the bag so the best we can do is try and use it responsibly. She discusses the “incredible destructive potential” of Gen AI on both the planet and our minds. Too many students are using AI to do the thinking for them rather than only engaging it as an augmentation of their own learning. Even in the early days of the internet I remember hearing myself saying “What do we need to remember things for when we can just look it up online whenever we need to?” I have seen the detriments to my own memory and critical thinking skills from the atrophy of thinking this way for so many years.
Links to video:
Learning with AI – Dr Mariel Miller
Additional Links
AI has and Environmental Problem – UN Environment Programme
From Plagiarism to Postplagiarism – Dr Sarah Eaton (Video)
Acceptable/Unacceptable – University of Alberta Student Life (Video)
European Network on Academic Integrity
Speculative Futures on ChatGPT and Generative Artificial Intelligence