In our 100th episode, TESOL Pop's co-founders Laura Wilkes and Eve Conway dive into the evolving world of generative AI and its implications for education. The explores challenges and opportunities tools like ChatGPT present in learning, assessment, and plagiarism.

Key talking points

What is Generative AI?
Eve breaks down the basics of generative AI and provides examples of what's currently available on the market, from text to art-generating tools. ChatGPT is highlighted as a tool that is having a growing influence on teaching and learning.

AI and Academic Integrity
The ease with which students can use AI to generate essays and assignments has raised concerns about plagiarism. Eve shares real examples of challenges teachers are facing now when it comes to coursework and assessment.

Future of Assessment
As AI continues to evolve, educators must rethink assessment methods. Eve and Laura discuss the shift towards externally assessed exams and question whether timed assessments are effective, especially for students needing extra support.

Classroom Strategies for AI Integration
Eve shares practical classroom strategies, such as comparing AI-generated texts with student work and using AI as a springboard for critical thinking. These strategies can help educators embrace AI while ensuring students develop their own ideas.

References & Resources

Links will direct you to resources Eve mentions in this episode and more.

1. Turnitin Plagiarism Detection Software.

   All hyperlinks listed were accessed on 03/09/2024.

Watch the video with closed captions or refer to the transcript below.

Transcript

00:00:00:08 - 00:00:19:03

Laura

TESOL Pop special episode 100th episode. Hi, everybody. Welcome to TESOL Pop; the mini podcast for busy. My name is Laura and joining me is fellow co-founder Eve Conway to talk about generative AI and the trends that we're seeing in education. Welcome back Eve. It's been a while.

00:00:19:05 - 00:00:27:02

Eve

Hi, Laura. Yeah, it's been a really long while. I'm really excited to be back for the 100th episode of TESOL Pop. Congratulations.

00:00:27:07 - 00:00:45:02

Laura

So I'm really glad we've got this time. And thanks for coming on the show to talk about this in terms of generative AI, like what we're talking about today. Just to kind of ground us in terms of what that looks like, what is generative AI and how is it kind of influencing, what we do as educators?

00:00:45:03 - 00:00:48:21

Laura

I think everybody knows about ChatGPT or maybe they don't. Actually, maybe we shouldn't assume.

00:00:48:23 - 00:01:09:17

Eve

Well, I think generative. If you break down the word it's AI that generates things. But that can be like many things, really. So the typical one we think of, the one that you met just mentioned was ChatGPT, where you enter a prompt and it writes a text so you can say, give me five recipes.

00:01:09:19 - 00:01:32:11

Eve

This this is what I use it for, five low carb recipes using these ingredients. And it will generate the recipes for you. But then there are so many uses of generative AI. It's not just limited to texts. So for example, artwork can be generated, you can brief an AI to create you an image. And even things like music and I think scripts to those.

00:01:32:13 - 00:01:45:17

Eve

I think a lot of controversy during LA writers strike about the use of AI in writing. I think a lot of actors and or script writers were worried that some of the work would be taken over by AI is to write those and produce those scripts.

00:01:45:17 - 00:02:03:11

Laura

I know you still work with educators. You work from an assessment perspective now, and examining perspective and creating exams and stuff. So I think between the two of us, we've been quite unique perspectives in what we're seeing for the classroom, for teachers and learners when it comes to generative AI. What are some of your thoughts on this?

00:02:03:12 - 00:02:21:06

Eve

Yeah, I mean, it's a big issue in education at the moment. I think all teachers have experienced students handing in work, which isn't their own work. Plagiarism is always been a big issue. And I think you can speak to any teacher about plagiarism, and they will have had that experience of the student submitting that work that isn't theirs.

00:02:21:06 - 00:02:41:13

Eve

But I think with AI now, it's just made it so easy for students to be able to do that. They can just type in a prompt and they can, get get a text written for them. Yeah. I was talking to one of my friends who still teaching a month or so ago, and, he was saying that his his students did it all the time sometimes.

00:02:41:13 - 00:03:18:04

Eve

But at the moment, anyway, I'm not sure whether AI is good enough to produce an essay that looks like it could have been written by a B1 student, or a B2 level student, because he said that he can always tell when his students have used AI to write an essay because they use, extremely complicated vocabulary. You know, like, this is a spellbinding classic which endures the test of time when when writing about a book, he said, I know that my B2 level student hasn't, written that themselves, that it's come from I.

00:03:18:06 - 00:03:39:04

Laura

Do you think that's because, though, that the students don't know how to prompt the AI yet quite yet, and they're still learning how to use it themselves, because you could instruct the AI to write. And I'm not I'm not trying to suggest on this podcast how to cheat in essays and stuff, but you could prompt it to say, could you use this type of vocabulary that would be appropriate for the CFR, level of a B2 student?

00:03:39:06 - 00:03:43:12

Laura

Within this context, with a bit more prompting, it could get better, couldn't it?

00:03:43:14 - 00:04:04:06

Eve

Yeah. I mean, it's learning all the time, isn't it? And is developing it all the time, and it's getting better all the time. So I would not be surprised if further down the line that's a possibility. I mean, it's always been that way, hasn't it, that there's a development, whatever the development is, and then people find ways with cheating to get around that development or around that rule.

00:04:04:06 - 00:04:34:18

Eve

Like when I was researching that today, I was looking at turn in, which is a popular AI detector. And, I just typed it into YouTube, to see what videos there were. And, one of the first things that came up on YouTube was, how to kind of get around, plagiarism detectors. So there are people out there making videos for students to learn how to circumnavigate plagiarism detector software already.

00:04:34:20 - 00:04:59:07

Laura

But this isn't something new, though, is it? Right? This has always existed in in the education industry. I remember when Wikipedia was quite new, and that's me showing my age in terms of how long we've been teaching, but how that kind of had an influence on what students were producing and how they used that. And I think to play the devil's advocate, some students didn't know how to use it as a resource and learning what the line is in terms of like using it as a research or to bounce ideas.

00:04:59:13 - 00:05:28:23

Laura

This is just kind of plagiarizing. And learning that was was definitely a gray area that we had to navigate. Thinking about like my early days teaching in, teaching in the college so that that's one aspect, isn't it? But I think you've always had, like, cheating as an issue in education beforehand. It used to be in the form of like, people could pay for essays like that was a whole industry probably still is where people can actually buy somebody time to write for them.

00:05:28:23 - 00:05:51:12

Laura

Of course, people could just copy from text and stuff like that. That's that was also something but a bit easier to detect. But I think with the former example of actually buying somebody's time, like a writer's time to write for you for essays and stuff like that, it was less accessible. It was very expensive. So what I feel with ChatGPT and Gemini and these like, text generating AI apps is that it's made it more accessible.

00:05:51:12 - 00:06:00:01

Laura

So it's now a bigger problem than what it was before, because it's almost like democratizing, access, having a private writer for you, you know?

00:06:00:03 - 00:06:07:00

Eve

Yeah, we like to democratize things in education that we can, and it feels like we've democratized cheating.

00:06:07:02 - 00:06:08:18

Laura

00:06:08:20 - 00:06:32:16

Eve

Yeah, it's a tough one. I mean, I think you're right. The cheating has always definitely been there. Like you mentioned, Wikipedia. I remember students using Google Translate, all the time to just translate, things from L1 straight into L2, which obviously would give you an awful piece of text, but they still did it anyway. The technology has got smarter in some ways.

00:06:32:16 - 00:07:08:18

Eve

So I think, like if I think back to my students using Google, Google Translate, it was so, so, so, so obvious that, this text was not written by a human because, or have been translated rather by AI because of the amount of mistakes in their inappropriate vocabulary. And I think with the AI written text, I think you can still tell that it wasn't written by the student, but it's often because it's so the language is so sophisticated and perhaps beyond the level.

00:07:08:20 - 00:07:28:17

Eve

I think the computers have gotten much smarter in the use of language, and I think just more sophisticated in their use of language. If you compare those texts that were churned out by by Google Translate and kind of had grammatical errors all over the place, you don't find that same level of grammatical inaccuracy in a ChatGPT text.

00:07:28:17 - 00:07:45:11

Eve

So I think that's the scary thing, is that they are getting better. And if they're getting better, then perhaps in the future a few steps down the line, like you said, they'll be able to write text that actually can imitate a student's writing quite accurately.

00:07:45:13 - 00:08:03:24

Laura

So I wonder what the way forward is, because there's obviously several use cases here like needs here that come to mind. One is like learner training and setting expectations for learners, like on how to use these tools because they're wonderful tools, right? When used well. But it can go the wrong way if they don't know what the guidelines are.

00:08:04:01 - 00:08:41:02

Laura

There's also guidelines that needed to be set by schools and training for teachers, so they feel confident in conveying that to the students and enacting guidelines if they need to. And there's also kind of the assessment itself, right. How do we assess students? I think that's where it's probably the most problematic that, you know, see, we're talking about in this conversation today is how can we assess students in a way that's fair, but also kind of proofs the assessment against easily being written by somebody else, whether that's an AI, somebody else, or you're paying somebody else to write it, because that has always been an issue.

00:08:41:04 - 00:09:20:13

Eve

It has. Yeah. There are a lot of big questions. I think in terms of like classroom usage and teaching, I think, teachers should talk to their students about AI. And they should make it clear, like, what's acceptable, what's not acceptable. My teacher friend that I mentioned before did a lesson on AI with his students, and he compared to text in the class like an AI written one and a student written one, and they looked for clues, to to show that the text was written by AI and the objectives of the lesson was for them to realize, just how stupid it was to plagiarize in that way, really, because

00:09:20:13 - 00:09:50:04

Eve

they were going to get found out. So I think that's an important discussion to have in the classroom. I also think that teachers could possibly explore using AI as prompts. So I did this one day with an essay that I was writing. I just wrote in a question and just looked at what came out. And I used it as a springboard to kind of think about my own ideas because I didn't actually like what the AI spat out.

00:09:50:04 - 00:10:06:21

Eve

I disagreed with most of the points that were on there. But it got me thinking about what my own ideas were. And then from that, I made a plan of my own ideas. And from that I was able to construct an essay. So I think that sort of usage could be useful.

00:10:06:23 - 00:10:36:23

Laura

Learner training, basically showing them how to use it. I think that's going to be an integrated part of just teaching, setting up assessments, the resources that you bring into the classroom. Right. I think this is brilliant. Like so guidelines involving the students in the guidelines, making it not just about guidelines and what your expectations are, but also supporting that critical thinking to help them also develop their digital literacy, AI literacy, if that could be a thing, and then also integrating it in as a tool, showing best practices and how to use it so that that's a really positive way of approaching it.

00:10:37:03 - 00:10:52:16

Laura

There was one other thing, wasn't it, that I was going to ask about it. And that's the assessments itself, because are we really going to go back to just getting students to sit in a room with a proctoring in place and somebody there to kind of do timed essays? Is that what we're going to end up going back to?

00:10:52:16 - 00:11:10:08

Laura

Because that doesn't feel so accessible? Oh, no. Quickly reliable. I remember some the exams I sat where I just felt awfully ill, and I had to sit there and still do it. And I know it wasn't my best, the best I could do. So yeah, it feels like that old model doesn't, you know, it doesn't have all the solutions either.

00:11:10:10 - 00:11:32:20

Eve

Yeah, it's a really tough one. Because I know that this is something that the Department for education, thinking about a lot is this use of AI and how to make sure that students work is their own work. I know that a lot of general qualifications now. So A-levels, GCSEs are now 100% externally assessed.

00:11:32:22 - 00:11:34:09

Laura

What does that mean?

00:11:34:11 - 00:12:05:16

Eve

Externally assessed means that students will complete an exam, and that exam will be marked by an examiner from the exam board, as opposed to doing, what we call a non exam assessment. That's more your traditional coursework options. So things like maybe students in the past when we were at school, we used to do essays for English as coursework, and our teacher would then mark them, and they'd just be moderated by the exam board.

00:12:05:17 - 00:12:14:10

Eve

A lot of subjects now for GCSE and A-level, the Department for education has set out that they have to be 100% externally assessed.

00:12:14:12 - 00:12:18:11

Laura

And does that mean also within time conditions as well for those externally assessed.

00:12:18:14 - 00:12:19:06

Eve

Yeah.

00:12:19:08 - 00:12:45:17

Laura

Pieces. Yeah that's that's really problematic isn't it. Because I think if we think about students that need that additional support, time flexibility for whatever their additional need may be, it then kind of puts them, at a disadvantage, doesn't it, in terms of their performance, because maybe they need extra time, maybe they need, support because of like dyslexia or something like that, or for health reasons or any other sort of reasons that are specific to that learners needs.

00:12:45:19 - 00:12:50:05

Laura

You know, we can't all perform that well. And time conditions on the specific day.

00:12:50:07 - 00:13:21:19

Eve

No, absolutely. And for certain subjects too, it just makes more sense to have coursework based, options. So, I mean, luckily, I work on a lot of vocational qualifications and, the, a lot of the vocational assessments that are out there at the moment still have, a lot of NEA non exam assessment, because it's recognized that practical work is such an important part of those qualifications.

00:13:21:21 - 00:13:37:14

Eve

But yeah, I think, if we move to a landscape where everything is 100% assessed, externally assessed, I think, like you say, the there will be issues with things like accessibility, and just assessment, validity.

00:13:37:14 - 00:13:54:04

Laura

It's difficult to know where it's going to go, isn't it. But I think it's, it's there's definitely needs to be structures in place, guidelines on how AI is used and maybe creative solutions and troubleshooting and what assessment is going to look like for the future. That kind of also doesn't exclude people.

00:13:54:06 - 00:13:56:03

Eve

Absolutely. Yeah I agree.

00:13:56:05 - 00:14:03:17

Laura

So overall, I'd like to close today's episode. How are you feeling about the future and I are you feeling quite optimistic?

00:14:03:19 - 00:14:31:20

Eve

I, I feel excited about it. I don't know if that's maybe, a false sense of excitement. I don't know whether I should be excited by it. I just think that whenever there's something new, there's also new opportunity. So it might be good to embrace some of those opportunities that are there in AI. But at the same time, be mindful of the risks that come with it.

00:14:32:01 - 00:14:41:23

Laura

Yeah, it's kind of like but having that balanced approach, isn't it? I can agree more if it's been lovely. Talk about this with you. Thanks so much for coming back on the show to chat about generative AI and assessment.

00:14:42:00 - 00:14:44:04

Eve

Thanks for having me. It's been great to be here.

00:14:44:10 - 00:15:00:24

Laura

So folks, what are your thoughts on this topic? Are you doing things in your school to embrace AI, or are you thinking of other things that you can do to assess your students? Share your ideas and thoughts on the website. We're going to use popcorn. You can find this particular episode and add your comments below. We'd love to hear from you.

00:15:01:02 - 00:15:16:18

Laura

Finally, you can support the work we do by leaving a rating, a review wherever you listen to the podcast, by sharing today's episode with your teaching community, or by even buying as a coffee, by going to ko-fi.com/tesolpop

You might also like

Good to know

This blog space and its associated multimedia content contain affiliate links. To learn more about how affiliate links work, please read our disclaimer here.