Ponderings on AI#
The rise of chatGPT has seen various ponderings, here's mine.
At some level the following three questions might be useful ways to talk about it.
Why might the types of thinking and discussion such questions might generate be important?
Other overview resources
What is the nature of the technology?#
This is not intended to be technological determinism (though it is a danger). The aim being to answer questions like
- How does this technology?
- What can it do?
- What are it is inherent limitations, requirements etc?
In the AI context, perhaps it is important to understand that
- They are not all knowing
- They maybe limited by their input
Related resources
- One look at our future - captures and discusses a Tweet giving some overview of the tech and its future
But also that there are different types of AI. The above are more specific to language models, the current AI of focus. There may be other types of AI.
What can we do with the technology? What are its affordances?#
Affordances is a useful concept because it is relational. Affordances arise from the relationship between the technology, user, their context, and purpose (task). i.e. not universal
Some related AI resources
- How ai can be used meaningfully by teachers and students in 2023 - USyd - good resource
- New modes of learning enabled by AI chatbots: Three methods and assignments
- Update your course syllabus for chatGPT
- Automated marking
- What is ChatGPT
- AI ChatGPT Overview
-
Tony Bates experimenting with ChatGPT and giving some thoughts
What's interesting is that the comments identify how Tony missed that the references (both websites and journal article citations) provided by ChatGPT were all fake. As pointed out in the comments, an example of how they nature of a generative language module makes this particular outcome more likely. It's aim is to make stuff that sounds probably. Not the right stuff. - Using it write a learning module on Grammar - act as a learning designer - use ChatGPT to develop a learning module in Liascript. - Three approaches - control; campaign; adoptive.
Academic misconduct - Thread on teacher detecting "academic misconduct" use of ChatGPT - GPTZero - tool using various attributes to determine if text was generated by AI
What will we do with these affordances? What does the critical perspective say?#
There's some blurring between this and the previous question. Where to draw the line between "can" and "will". Perhaps the focus here is more on the critical take on what the contexts we operate in will allow/encourage/enable. But it's also about people themselves.
Perhaps a good example is the differences between John Perry Barlow's A Declaration of the Independence of Cyberspace - as an indicator of the optimism around the promise of the Internet - and the reality of the contemporary era of fake news, influencers and social media. Arguably demonstrating the "can"/"will" difference.
Some related AI resources (some of which blur between can/will)