|
Getting your Trinity Audio player ready...
|
As a teacher, you are constantly looking for new and innovative ways to engage your students and improve their learning experience. One tool that you may not have considered is ChatGPT, a large language model developed by OpenAI. In this blog post, we will explore three ways that teachers can use ChatGPT to enhance teaching and learning in the classroom.
- Facilitate discussions and promote critical thinking: ChatGPT can help facilitate discussions and promote critical thinking by providing suggestions and prompts for students to consider. For example, you can use ChatGPT to generate a list of open-ended questions related to a particular topic, and then have students discuss and respond to these questions in small groups or as a whole class. This can help students develop their critical thinking skills and deepen their understanding of the topic at hand.
- Provide personalized feedback and support: ChatGPT can also be used to provide personalized feedback and support to individual students. For example, you can use the tool to generate specific feedback on a student’s writing assignment, providing suggestions for improvement and encouragement along the way. This can help students feel more confident and motivated to continue learning and growing.
- Enhance language learning and literacy skills: ChatGPT can be a valuable tool for enhancing language learning and literacy skills. For students learning a second language, ChatGPT can provide instant feedback on their pronunciation and grammar, helping them improve their language skills quickly and efficiently. For students who are struggling with reading and writing, ChatGPT can provide support by generating suggestions for improving their sentences and paragraphs.
In conclusion, ChatGPT is a powerful tool that can help teachers enhance teaching and learning in the classroom. By using ChatGPT to facilitate discussions, provide personalized feedback, and enhance language learning and literacy skills, teachers can create a more engaging and effective learning environment for their students.
Ok flesh and blood human here. Yes, the above is an auto-generated article using ChatGPT.
I was tempted to just post this as is without any additional commentary, but then thought I needed to add something about how I am just playing around with both Dall-e, which auto-generated the image based on the prompt in the caption (which, as an aside, is doing double duty as I have also used it as the alt text for the image) and ChatGPT, which auto-generated the actual article.
Honestly, I am struck by how good this bot-written post is. The post is missing one of the elements that I think make a good blog post and that is a reflection of the personality of the author, so it does feel a bit bland and generic. But as I read this I thought to myself “oh those aren’t bad suggestions.”
I am not sure where I am landing on the whole AI auto-generated content just yet. There is a part of me that thinks back to a kinder, gentler more naive time in the technology history timeline and thinks that this is just a natural outcome of the talk we once had about ways in which technology could be used to harness our collective intelligence. There is a part of me that is quite excited by the possibilities of this technology, although I know, from an education perspective, this is going to drive a lot of traditionalists around the bend with worries of academic integrity and mass cheating, although I think (hope) that the accessibility of these types of tools may actually spur a closer look at traditional assessment practices. Sadly though, I think the reality will be educators doubling down on surveillance technology, virtual proctoring, browser lock-downs and the like rather than doing the deep and hard work of re-evaluating assessment practices.
When I use these tools, something I do feel is a sense of creativity, especially with Dall-e. I have spent quite a bit of time with Dall-e trying to create a descriptive prompt that will generate an image that comes close to what I have in my head. That feels like a creative process. A way to translate a vision I have into words and have a tool create something I never could create without that tool. And it feels creatively empowering.
There are, of course, numerous issues around AI-generated content – its use & misuse for mis/disinformation, the ethics around the datasets used to train these systems, and how to attribute works that have been created with these tools. So there are many reasons to be critical and keep my optimism grounded.
The one thing I feel certain about is that we have crossed into some new territory with the OpenAI tools, and that one day we may look back on 2022 as the year that AI-generated content went mainstream. Yes, AI-generated content is nothing new, but Dall-e and ChatGPT are tools that are easy enough to use that AI-generated content is now possible by almost anyone who can write a prompt in a box. The killer app’s of AI-genertaed content. And that feels significant.
7 Comments
So did ChatGPT write this?
Yes it did. I just went to ChatGPT, entered your blog post as a title, and it spewed out the exact same text.
Now this makes me curious, is ChatGPT storing prompt/response pairs? Does it always respond the same?
And lastly, you did not attribute your co-author 😉
Interesting. I suspect you are right and it is storing responses to preserve computing cycles. I wonder if that is a default behaviour or something that OpenAI might be doing while it is in an open beta to preserve computing power for redundant prompts?
Was it not clear in the post what was written/created by ChatGPT and DALL-E? I was worried about that so tried to be extra explicit in the post where ChatGPT ended and I began.
You were clear, but at some point maybe being able to make this distinction in text form will get more challenging.
I wonder if AI will learn to respone with allmy typoes and bad kayboarding skills? 😉
The attribution bit does take on a different dimension when we talk about attributing AI content vs human content, doesn’t it? I mean, attribution of another person’s work is a recognition that that person has influenced your thinking, or a respectful recognition of their work, or even a simple thank you (as is how I think of CC attributions most of the time). But with AI content attribution will have a different dimension in that it will be an important notifier to the reader/viewer that the content was not created by a person. Attributing to AI sends a different message to the viewer than attributing to another person. At least that is how I am thinking about an attribution requirement when it comes to AI content – a way to notify the person reading that the content was AI-generated.
I guess at a higher level there is also a moral case to be made that these algorithms are often trained by corpi created by people and that those people who contributed content to train the AI be attributed. But at some stage you have to wonder how practical it is to even do that.
I’m also the same as you re) typos & spleling over here. You can always tell I, an actual human with all my grammatical flaws and foibles, have written something because there will be at least one typo DESPITE having no shortage of technology tools at hand to help with my spelling.
Clint, Thanks for this. Made me look at some of these tools. What are the licensing implications of the generated images with Dall-e? Does it use CC images as the base? What would be the attribution of a generated image?
Rosario
Rosario, I’ve been wondering the same. Less so about the source images, I think the way images are used as training is much different to our intuition of what remix/derivative works are. The licensing is murky too, some legal minds suggest media created my a non human cannot be copyrighted (hence all is public domain?), but there is no legal determination on that.
I have been asking more about the ways we attribute these images, especially as DALL-E has been regularly changing their terms of service, and until late there was no source for them. It’s still murky, I started some discussions here
https://connect.oeglobal.org/t/the-grey-zone-use-attribution-of-statistical-generated-imagery/4328
and Creative Commons has recently published some info on their policy efforts in AI
https://creativecommons.org/2022/12/03/our-work-in-policy-at-cc-artificial-intelligence/
It’s a great question, Rosarrio, and one that people in the open community are starting to grapple with as it isn’t really clear just yet. Alan Levine wrote a blog post last month that started to dig into some of the nuances of licensing and attribution https://cogdogblog.com/2022/11/hello-dall-e/ and Alan has also been posting about this very issue in the OE Global forums https://connect.oeglobal.org/t/the-grey-zone-use-attribution-of-statistical-generated-imagery/4328