WEBVTT 00:00:08.300 --> 00:00:11.839 Opening AI for language learning explores artificial 00:00:11.839 --> 00:00:14.500 intelligence through informal conversations between 00:00:14.500 --> 00:00:17.019 two longtime colleagues in language education 00:00:17.019 --> 00:00:20.940 and digital technologies, Mat Schulze and Phil 00:00:20.940 --> 00:00:25.359 Hubbard. Language educators, learners, and enthusiasts 00:00:25.359 --> 00:00:28.160 are invited to an informed discussion where emerging 00:00:28.160 --> 00:00:34.340 technology meets human communication. Well, Mat, 00:00:34.359 --> 00:00:40.619 here we are back again. Yes, Phil. What's on 00:00:40.619 --> 00:00:43.740 the menu today? What are we doing? Well, I'd 00:00:43.740 --> 00:00:49.280 like to talk about a specific kind of GenAI SIPD, 00:00:49.539 --> 00:00:54.020 going back to our paper that started this, the 00:00:54.020 --> 00:00:56.979 generative AI sustained integrated professional 00:00:56.979 --> 00:01:01.899 development. And this is one of our recommendations, 00:01:02.060 --> 00:01:08.189 which is, you know, do it. Somewhat collaboratively 00:01:08.189 --> 00:01:12.090 don't do everything by yourself So I engaged 00:01:12.090 --> 00:01:18.129 in one such thing recently Sounds good Seymour 00:01:18.129 --> 00:01:23.750 I went to a conference So it was TESOL 00:01:23.750 --> 00:01:29.030 2026 in Salt Lake City from March 24th to 27th 00:01:29.030 --> 00:01:32.769 I was giving a presentation there, but mostly 00:01:32.769 --> 00:01:36.819 I wanted to see what practice -focused organization 00:01:36.819 --> 00:01:41.519 like TESOL was doing with AI, because a lot of 00:01:41.519 --> 00:01:44.700 the conferences we go to are either mixed research 00:01:44.700 --> 00:01:48.280 and practice with a heavy emphasis on research 00:01:48.280 --> 00:01:52.680 or really almost pure research, even though it 00:01:52.680 --> 00:01:58.780 may be applied research. That's an interesting 00:01:58.780 --> 00:02:02.590 one coming basically from another angle. What 00:02:02.590 --> 00:02:04.909 are they doing? What's happening with AI and 00:02:04.909 --> 00:02:10.810 TESOL? Well, I didn't go last year, so this is 00:02:10.810 --> 00:02:12.949 kind of catching up. I can't say what happened 00:02:12.949 --> 00:02:16.409 in 2025, but now they seem to be doing a lot 00:02:16.409 --> 00:02:21.629 more than they were doing in 2024. I arrived 00:02:21.629 --> 00:02:25.870 at the conference before it began, so I had the 00:02:25.870 --> 00:02:30.750 entire time. My presentation had been... nicely 00:02:30.750 --> 00:02:35.150 placed on the last day in the last slot. And 00:02:35.150 --> 00:02:37.569 because I was doing it with a colleague, Deborah 00:02:37.569 --> 00:02:40.909 Healy, we had to have everything organized in 00:02:40.909 --> 00:02:43.830 advance so I couldn't be working on my slides 00:02:43.830 --> 00:02:49.509 instead of going to the conference. And that 00:02:49.509 --> 00:02:53.110 meant I had a lot of time to do this. So I went 00:02:53.110 --> 00:02:57.069 to as many AI sessions of one sort or another 00:02:57.069 --> 00:02:59.849 as I could. find, and often there was more than 00:02:59.849 --> 00:03:02.449 one at the same time, so I had to make choices, 00:03:03.569 --> 00:03:11.530 presentations and panels and posters, but I also 00:03:11.530 --> 00:03:15.430 checked out relevant exhibitors in the exhibit 00:03:15.430 --> 00:03:18.550 hall, a couple of which I'd known about before, 00:03:18.550 --> 00:03:21.550 and then the interest section meetings, especially 00:03:21.550 --> 00:03:24.650 for the call interest section and the teacher 00:03:24.650 --> 00:03:29.719 ed interest section. I also visited the TESOL 00:03:29.719 --> 00:03:32.620 publications booth and I'll say a little bit 00:03:32.620 --> 00:03:35.520 more on that later, but they've got some publications 00:03:35.520 --> 00:03:41.979 out. Nice. This does sound like a lot. So overall, 00:03:42.159 --> 00:03:46.360 what was your impression? What did you kind of 00:03:46.360 --> 00:03:50.979 take home, as they say? Well, I have a lot of 00:03:50.979 --> 00:03:55.199 notes and maybe half of them I'm actually able 00:03:55.199 --> 00:04:02.009 to decipher. One of them actually looked like 00:04:02.009 --> 00:04:04.810 it was written with the Enigma machine. But anyway, 00:04:05.770 --> 00:04:09.949 I also took photos of slides and picked up a 00:04:09.949 --> 00:04:12.990 lot of things. So there was way too much that 00:04:12.990 --> 00:04:16.329 we have time to talk about here, but some general 00:04:16.329 --> 00:04:21.350 takeaways, interesting balance between skepticism 00:04:21.350 --> 00:04:24.720 and excitement. So there were some presentations 00:04:24.720 --> 00:04:28.240 that were, you know, were not jumping into the 00:04:28.240 --> 00:04:31.839 hype. There was some hype, of course, and some 00:04:31.839 --> 00:04:35.660 what I would think of as naive application people 00:04:35.660 --> 00:04:39.199 who had recently discovered this and were presenting 00:04:39.199 --> 00:04:41.199 it, but there were teachers in the audience who 00:04:41.199 --> 00:04:47.740 that was appropriate for. A lot on ethics, sometimes 00:04:47.740 --> 00:04:50.259 as part of the talk, but sometimes the main thing. 00:04:50.420 --> 00:04:55.100 There was a whole call intersection panel on 00:04:55.100 --> 00:05:01.399 AI ethics, and also a lot on critical AI literacy, 00:05:02.079 --> 00:05:05.800 going beyond just saying, well, here's what AI 00:05:05.800 --> 00:05:09.319 is and here's how you use it, but some of the 00:05:09.319 --> 00:05:14.600 things that we've talked about before. And commercial 00:05:14.600 --> 00:05:17.259 developers, where they're in the exhibit hall, 00:05:17.680 --> 00:05:23.060 they were mostly small -scale startups, as opposed 00:05:23.060 --> 00:05:27.060 to the giant players that used to show up back 00:05:27.060 --> 00:05:29.740 in the era of CD -ROMs where you had to spend 00:05:29.740 --> 00:05:34.920 multi -million dollars on projects in order to 00:05:34.920 --> 00:05:39.220 have a whole set of multimedia materials to go. 00:05:40.000 --> 00:05:42.920 Duolingo was there, so certainly a big player 00:05:42.920 --> 00:05:46.509 there. But a lot of these were focused on what 00:05:46.509 --> 00:05:50.209 I think of as school solutions, which made sense. 00:05:52.189 --> 00:05:56.470 A couple of them do have materials that can be 00:05:56.470 --> 00:05:59.410 used well independently, but in a lot of cases, 00:05:59.550 --> 00:06:05.430 they're trying to bring in schools or even school 00:06:05.430 --> 00:06:09.560 districts language. language organizations of 00:06:09.560 --> 00:06:13.139 various sorts and try to sell whole solutions 00:06:13.139 --> 00:06:16.000 to them either to support their curriculum or 00:06:16.000 --> 00:06:19.899 even embed the curriculum. So again, I'll say 00:06:19.899 --> 00:06:25.240 more about that in a bit. Great. Yeah, this a 00:06:25.240 --> 00:06:29.420 lot seems to have been going on at the TESOL 00:06:29.420 --> 00:06:33.220 conference. You said earlier the year before 00:06:33.220 --> 00:06:37.560 you didn't go, but you went in in 2004. I'm sure 00:06:37.560 --> 00:06:40.660 you've been going to TESOL quite regularly. You 00:06:40.660 --> 00:06:43.920 must have had some expectations looking at this 00:06:43.920 --> 00:06:47.100 AI in particular. So what were your expectations 00:06:47.100 --> 00:06:51.759 and where they met basically from your perspective? 00:06:52.339 --> 00:06:55.120 Well, it was even though I had been going to 00:06:55.120 --> 00:07:02.370 TESOL almost every year since 1984, This was 00:07:02.370 --> 00:07:05.310 different for me. My experience was different 00:07:05.310 --> 00:07:09.290 from previous conferences. In what way? How was 00:07:09.290 --> 00:07:15.110 it different? Well, for years, I've gone to conferences. 00:07:15.250 --> 00:07:17.490 I've always been interested in focusing on the 00:07:17.490 --> 00:07:23.930 technology stuff. But I was going in the, I think, 00:07:24.589 --> 00:07:27.269 in the position of someone who already knew a 00:07:27.269 --> 00:07:31.339 whole lot and was even deemed an expert in particular 00:07:31.339 --> 00:07:36.899 elements of call. I joined TESOL call intersection 00:07:36.899 --> 00:07:40.939 in its first year in 1984, and I've grown up 00:07:40.939 --> 00:07:47.079 with the field. We've certainly seen lots of 00:07:47.079 --> 00:07:50.860 changes in the last 40 plus years, but the changes 00:07:50.860 --> 00:07:55.259 have not been so dramatic and so rapid as with 00:07:55.259 --> 00:08:00.370 AI. For me, instead of you know, just hanging 00:08:00.370 --> 00:08:03.470 out with friends. Teesaw used to have what they 00:08:03.470 --> 00:08:05.529 call the electronic village and I would tend 00:08:05.529 --> 00:08:07.930 to camp in there and volunteer to help people 00:08:07.930 --> 00:08:11.430 out and things like that. Now I'm going to sessions, 00:08:11.889 --> 00:08:14.470 kind of like I'm a graduate student all over 00:08:14.470 --> 00:08:17.810 again and trying to learn as much as I can during 00:08:17.810 --> 00:08:24.930 these four days. And what I discovered really 00:08:24.930 --> 00:08:27.550 quickly, I already knew going in is that I'm 00:08:27.550 --> 00:08:31.649 not an AI expert. I can consider myself a call 00:08:31.649 --> 00:08:34.909 expert in a lot of areas anyway, but I'm not 00:08:34.909 --> 00:08:38.830 an AI expert. One of the talks, in fact, that 00:08:38.830 --> 00:08:43.649 I went to that I'll mention later, Jane Freeman 00:08:43.649 --> 00:08:46.789 from the University of Toronto basically said, 00:08:46.789 --> 00:08:51.129 none of us are AI literate, and then proceeded 00:08:51.129 --> 00:08:55.590 to give a really good talk that showed that she 00:08:55.590 --> 00:09:01.809 knows a lot. So I think she's right. I don't 00:09:01.809 --> 00:09:04.470 know her and obviously I wasn't at that presentation, 00:09:04.549 --> 00:09:11.590 but I agree. We're not too literate about things. 00:09:11.909 --> 00:09:15.690 It's very much a learning process. So what are 00:09:15.690 --> 00:09:17.809 we going to do about that then if we are these 00:09:17.809 --> 00:09:22.419 kind of semi -literate people? We can always 00:09:22.419 --> 00:09:25.240 fall back on how we think we know more than we 00:09:25.240 --> 00:09:32.919 do. I still think I know more than I do. And 00:09:32.919 --> 00:09:35.600 of course, how we don't know what we don't know. 00:09:36.080 --> 00:09:38.700 So there's that. That's not very productive. 00:09:39.200 --> 00:09:42.059 I actually want to be AI literate, whatever that 00:09:42.059 --> 00:09:46.440 means, both in general and for the specific purpose 00:09:46.440 --> 00:09:51.610 of language education. I hear you. I think it's 00:09:51.610 --> 00:09:54.830 a challenge we all face. And yes, it has a lot 00:09:54.830 --> 00:09:57.669 to do with motivation, right? The interest in 00:09:57.669 --> 00:10:00.450 learning about AI, which is something the two 00:10:00.450 --> 00:10:04.789 of us certainly share. And I think, yeah, that's 00:10:04.789 --> 00:10:07.129 the main thing one can ask for. It's an openness, 00:10:07.470 --> 00:10:10.570 it's a willingness to learn. And then, yeah. 00:10:12.129 --> 00:10:16.830 So thinking about this AI literacy, it gets thrown 00:10:16.830 --> 00:10:21.059 around a lot. I've heard that term, the concept, 00:10:21.200 --> 00:10:23.759 the phrase very often. It's almost becoming, 00:10:24.700 --> 00:10:26.700 from my personal perspective, a little bit hollow 00:10:26.700 --> 00:10:30.100 because it's used so often. And it seems to be 00:10:30.100 --> 00:10:34.240 used in many different ways, which means it loses 00:10:34.240 --> 00:10:38.720 some of its meaning. Speaking personally, I would 00:10:38.720 --> 00:10:41.980 say I'm not really sure what AI literacy actually 00:10:41.980 --> 00:10:45.259 means in very practical terms. What am I going 00:10:45.259 --> 00:10:47.659 to do about it? If I want to have it, basically, 00:10:47.679 --> 00:10:53.769 what do I need? So I think it is important to 00:10:53.769 --> 00:10:57.429 poke a couple of holes into the things we think 00:10:57.429 --> 00:11:01.210 we know. Because this is the only way to make 00:11:01.210 --> 00:11:05.730 sure that we actually at least know where the 00:11:05.730 --> 00:11:09.690 gap is for us. Where are the things that we don't 00:11:09.690 --> 00:11:14.840 yet know or maybe can't know. This reminded me 00:11:14.840 --> 00:11:17.059 of this phrase I heard very often when going 00:11:17.059 --> 00:11:19.720 to London, England, basically in the London Underground, 00:11:19.919 --> 00:11:22.159 there was always this disembodied voice telling 00:11:22.159 --> 00:11:26.799 you, mind the gap. The doors are opening, mind 00:11:26.799 --> 00:11:32.919 the gap. So what are examples of some specific 00:11:32.919 --> 00:11:36.340 experiences you had at TESOL, where you probably 00:11:36.340 --> 00:11:38.679 minded the gap also, where you learned something 00:11:38.679 --> 00:11:46.059 new? Well, at first, facing what to share, I 00:11:46.059 --> 00:11:49.899 thought, oh, I'll give a lot of different examples 00:11:49.899 --> 00:11:52.700 and say a little bit about each one. And I realized 00:11:52.700 --> 00:11:57.779 that that just doesn't give people as much to 00:11:57.779 --> 00:11:59.620 take away. They might as well just look at the 00:11:59.620 --> 00:12:03.840 list of abstracts. So instead, I'm going to talk 00:12:03.840 --> 00:12:09.580 about a few areas that had an impact on me for 00:12:09.580 --> 00:12:13.600 one reason or another. and that I think represents 00:12:13.600 --> 00:12:16.159 something of the range of what was there. So 00:12:16.159 --> 00:12:21.840 the first example by a presenter named Miriam 00:12:21.840 --> 00:12:26.460 Moore, who was offering a four -part critical 00:12:26.460 --> 00:12:32.379 AI literacy framework. And the four parts, well, 00:12:32.419 --> 00:12:35.879 first of all, this is, I think, embedded in a 00:12:35.879 --> 00:12:39.769 pedagogy. class of hers, although it's also available 00:12:39.769 --> 00:12:45.169 in more workshopy kind of format for teachers 00:12:45.169 --> 00:12:46.929 to use for their own professional development. 00:12:47.590 --> 00:12:50.629 So it starts with, not surprisingly the way we 00:12:50.629 --> 00:12:55.669 did, with what is AI? But then it goes to, instead 00:12:55.669 --> 00:13:01.429 of saying, OK, ethics or chat bots or translation, 00:13:02.570 --> 00:13:06.070 the second part is, what are my rights and responsibilities? 00:13:06.700 --> 00:13:10.200 That's a good one. The third part, who makes 00:13:10.200 --> 00:13:14.840 the rules for AI? And then the fourth one, how 00:13:14.840 --> 00:13:18.840 does the language about AI influence our beliefs 00:13:18.840 --> 00:13:22.159 and how we use it? This is an Emily Bender sort 00:13:22.159 --> 00:13:30.460 of thing. And interestingly, she mentioned the 00:13:30.460 --> 00:13:35.129 Mystery Hype Theater AI Hype Theater 3000. and 00:13:35.129 --> 00:13:40.590 also introduced Usta to a book by Karen Howe 00:13:40.590 --> 00:13:44.990 called Empire of AI. So she reads and interprets 00:13:44.990 --> 00:13:49.629 in these areas, but she's not trying to say, 00:13:50.169 --> 00:13:55.970 don't use AI. She's trying to say, here is what 00:13:55.970 --> 00:14:03.899 it means to be critically AI literate. She had 00:14:03.899 --> 00:14:11.480 this really interesting demo explaining how a 00:14:11.480 --> 00:14:15.100 chat bot works and how a large language model 00:14:15.100 --> 00:14:17.799 works. She asked for six volunteers from the 00:14:17.799 --> 00:14:22.519 audience and she had them get up. Then she said, 00:14:22.759 --> 00:14:25.960 I've got you in this order and each one of you 00:14:25.960 --> 00:14:30.159 is going to get to say one word. the first in 00:14:30.159 --> 00:14:32.779 answer to a question I'll ask and then the next 00:14:32.779 --> 00:14:35.559 person has to put a word in that makes sense 00:14:35.559 --> 00:14:37.860 following that first word and also makes sense 00:14:37.860 --> 00:14:43.179 for the question and so the question was What 00:14:43.179 --> 00:14:47.340 is the FIFA World Cup and the first person said 00:14:47.340 --> 00:14:52.679 it and the next person is and then a and then 00:14:52.679 --> 00:14:57.750 soccer tournament And the sixth one was, it was 00:14:57.750 --> 00:14:59.529 kind of at the end, the sixth one had to come 00:14:59.529 --> 00:15:02.889 up with something and said, worldwide. And so 00:15:02.889 --> 00:15:07.629 it was, you know, what's a probable next word? 00:15:09.090 --> 00:15:14.830 That's what it's doing, you know. It's text extruding 00:15:14.830 --> 00:15:18.389 machines. Anyway. It does it so much more quickly 00:15:18.389 --> 00:15:21.230 than six people can, you know. Yeah. Well, yeah, 00:15:21.230 --> 00:15:24.250 but I think it was a beautiful demonstration. 00:15:24.509 --> 00:15:27.500 Yeah. It's a very neat idea. You know, simple, 00:15:27.620 --> 00:15:31.440 but also quite engaging, I'm sure for the volunteers, 00:15:31.580 --> 00:15:34.940 but for the audience as well. So there were a 00:15:34.940 --> 00:15:38.299 number of useful professional development ideas, 00:15:38.320 --> 00:15:41.740 a couple of ones that stuck out for me and that 00:15:41.740 --> 00:15:46.320 I probably should do more myself. One was instead 00:15:46.320 --> 00:15:51.100 of just studying ethics or even looking at, you 00:15:51.100 --> 00:15:53.120 know, what is your institutional stuff? How do 00:15:53.120 --> 00:15:55.950 you figure it out? figure out what to tell your 00:15:55.950 --> 00:15:59.269 students. Start with something like a personal 00:15:59.269 --> 00:16:03.049 ethical checklist. So what does it really mean 00:16:03.049 --> 00:16:05.309 for you? It seems to me there's going to be a 00:16:05.309 --> 00:16:07.529 lot of thought that has to go into that. And 00:16:07.529 --> 00:16:11.730 then something I've done piecemeal but not systematically 00:16:11.730 --> 00:16:16.850 is conduct a privacy audit of all the AI platforms 00:16:16.850 --> 00:16:21.169 you use and see exactly what's in those privacy 00:16:21.169 --> 00:16:23.169 statements, what you have control of, what you 00:16:23.169 --> 00:16:26.250 don't have control of. Oh, the privacy statements. 00:16:26.350 --> 00:16:28.110 Yeah, I was wondering how you would do an audit 00:16:28.110 --> 00:16:30.570 basically, but yeah, read the privacy statements. 00:16:30.769 --> 00:16:34.909 That's a very good first step. And also notice 00:16:34.909 --> 00:16:37.470 what's not there potentially that should be. 00:16:38.429 --> 00:16:43.509 OK, so that's the first example. The second one, 00:16:43.509 --> 00:16:45.789 this is the one I mentioned earlier, Jane Freeman 00:16:45.789 --> 00:16:49.169 from University of Toronto. This was part of 00:16:49.169 --> 00:16:55.000 a panel on AI for program administrators. And 00:16:55.000 --> 00:16:58.620 she was talking in general about the development 00:16:58.620 --> 00:17:02.659 of expertise. And as I noted earlier, she was 00:17:02.659 --> 00:17:05.980 the one who quoted, none of us are AI literate, 00:17:06.200 --> 00:17:08.960 that this expertise is something evolving as 00:17:08.960 --> 00:17:12.400 the AI evolves. And it really dovetails nicely 00:17:12.400 --> 00:17:16.339 with our sustained integrated professional development 00:17:16.339 --> 00:17:21.140 idea, especially the sustained part of it. So 00:17:21.140 --> 00:17:25.769 she gave an example from from medicine using 00:17:25.769 --> 00:17:30.910 this analogy of puppet versus puppeteer and said, 00:17:31.069 --> 00:17:34.029 okay, if you've got an undergraduate who wants 00:17:34.029 --> 00:17:36.769 to be a doctor. So a first year undergraduate 00:17:36.769 --> 00:17:40.349 is basically a puppet in the sense of they don't 00:17:40.349 --> 00:17:44.329 have a lot of agency. They do what the instructor 00:17:44.329 --> 00:17:48.589 tells them to do. They work on things because 00:17:48.589 --> 00:17:50.799 they don't have. the knowledge yet, and they 00:17:50.799 --> 00:17:53.720 don't have the skills. And at the top, the puppeteer 00:17:53.720 --> 00:17:56.480 is ultimately like the full -fledged doctor, 00:17:58.519 --> 00:18:02.339 the attending physician, for example. And then 00:18:02.339 --> 00:18:04.859 various levels in between of the first -year 00:18:04.859 --> 00:18:07.420 medical student is above the first -year graduate 00:18:07.420 --> 00:18:11.000 and so on. And it all seems to have to do with 00:18:11.000 --> 00:18:17.559 this locus of control. And then she applies this 00:18:17.559 --> 00:18:24.559 to AI. and says, okay, we can show people can 00:18:24.559 --> 00:18:28.720 immediately make use of ChatGPT, for example, 00:18:29.079 --> 00:18:31.880 or any of the other ones to get a question answered 00:18:31.880 --> 00:18:34.579 and to produce something. You've mentioned before, 00:18:34.740 --> 00:18:38.220 it's like you can ask it to produce a lesson 00:18:38.220 --> 00:18:41.880 plan before you know anything about how to produce 00:18:41.880 --> 00:18:46.099 a lesson plan yourself. And that's the idea. 00:18:47.069 --> 00:18:50.049 The AI is the puppeteer. It's telling you what 00:18:50.049 --> 00:18:55.349 you should put into a lesson plan. But it's not 00:18:55.349 --> 00:18:58.569 a person. It's not a skilled teacher, as you 00:18:58.569 --> 00:19:02.490 point out. It's drawing on various resources 00:19:02.490 --> 00:19:08.809 to put something coherent seeming together. And 00:19:08.809 --> 00:19:13.569 so she uses the term, the two sort of opposite 00:19:13.569 --> 00:19:18.970 terms are cognitive offloading. and distributed 00:19:18.970 --> 00:19:24.269 cognition. And if you're engaged in cognitive 00:19:24.269 --> 00:19:28.710 offloading, for example, you're trying to learn 00:19:28.710 --> 00:19:34.990 how to write a particular kind of email. Let's 00:19:34.990 --> 00:19:40.410 say you have an email and you're asking someone 00:19:40.410 --> 00:19:44.819 to sponsor your research, for example. And you 00:19:44.819 --> 00:19:47.500 don't know how to do that. So you ask CHAP GPT 00:19:47.500 --> 00:19:50.099 to write it for you. And then you look at it 00:19:50.099 --> 00:19:52.099 and you go, oh, that's okay. And you send it 00:19:52.099 --> 00:19:56.000 out. You haven't learned anything. It doesn't 00:19:56.000 --> 00:19:58.599 directly represent you. You're the puppet there. 00:20:00.619 --> 00:20:03.920 And if you're cognitive offloading and you're 00:20:03.920 --> 00:20:08.259 already an expert at something and you can glance 00:20:08.259 --> 00:20:10.240 at the output and say, okay, that's good. And 00:20:10.240 --> 00:20:14.970 you can also put in better input. then you're 00:20:14.970 --> 00:20:16.869 in a better position. So it's the same thing 00:20:16.869 --> 00:20:19.710 with use of AI is you want to move from being 00:20:19.710 --> 00:20:23.470 the puppet to being the puppeteer. And you can 00:20:23.470 --> 00:20:27.329 only be the puppeteer if you already have significant 00:20:27.329 --> 00:20:32.309 domain knowledge and skill. And this is anyway, 00:20:32.589 --> 00:20:36.589 that's if I interpreted her work correctly, it 00:20:36.589 --> 00:20:42.670 comes down to learning how to use AI in a way 00:20:42.670 --> 00:20:46.099 that while it supports student and teacher agency, 00:20:46.599 --> 00:20:54.960 it's agency in areas where they can have some 00:20:54.960 --> 00:20:58.480 engagement with distributed cognition and not 00:20:58.480 --> 00:21:05.920 just cognitive offloading. This whole concept 00:21:05.920 --> 00:21:09.640 of cognitive offloading is very interesting. 00:21:09.740 --> 00:21:13.809 We should probably put a pin in and think about 00:21:13.809 --> 00:21:16.609 that one again. Because there's also, of course, 00:21:17.470 --> 00:21:19.490 positive cognitive offloading, right? That helps 00:21:19.490 --> 00:21:24.049 learning if I offload stuff that I don't really 00:21:24.049 --> 00:21:26.690 need to do again in order to learn something 00:21:26.690 --> 00:21:29.509 new. Yeah. And I think part of this, too, had 00:21:29.509 --> 00:21:34.349 to do with it was like a lot of models in order 00:21:34.349 --> 00:21:39.250 to be simple and clear. go back and poke holes 00:21:39.250 --> 00:21:41.670 in it and say, okay, what it really comes down 00:21:41.670 --> 00:21:47.869 to is that if you are a novice, that you're trying 00:21:47.869 --> 00:21:52.369 to, let's take the lesson plan example. You could 00:21:52.369 --> 00:21:56.869 actually use examples of lesson plans produced 00:21:56.869 --> 00:22:01.849 by AI to help and deconstruct them and say, oh, 00:22:01.930 --> 00:22:04.529 I see here like four different lesson plans. 00:22:04.970 --> 00:22:08.769 What do they have in common? that need to go 00:22:08.769 --> 00:22:11.410 into lesson plans. And how does that compare 00:22:11.410 --> 00:22:14.750 to a teacher -produced lesson plan that, you 00:22:14.750 --> 00:22:18.690 know, my instructor brought up in class? And 00:22:18.690 --> 00:22:22.029 so those, that's where I think the interaction 00:22:22.029 --> 00:22:28.730 could be useful. Okay. The third area, and this 00:22:28.730 --> 00:22:33.450 involves some multiple characters, is the exhibitors. 00:22:34.029 --> 00:22:39.529 And as I mentioned, there were several more or 00:22:39.529 --> 00:22:44.730 less startups, I think, as well as Duolingo is 00:22:44.730 --> 00:22:51.750 the 800 pound gorilla in the room. And then other 00:22:51.750 --> 00:22:54.930 publishers like Pearson, for example, doing things 00:22:54.930 --> 00:22:58.089 with AI, but always connected to what they already 00:22:58.089 --> 00:23:02.289 have as essentially a textbook driven infrastructure. 00:23:04.170 --> 00:23:08.450 So one of these, actually two of them were ones 00:23:08.450 --> 00:23:12.490 that I had touched base with briefly at Calico 00:23:12.490 --> 00:23:17.690 last year. One of them is called FlowSpeak. I 00:23:17.690 --> 00:23:21.490 talked to the developer Matt Sussman there and 00:23:21.490 --> 00:23:25.950 FlowSpeak has basically semi -structured scenarios 00:23:25.950 --> 00:23:32.670 where you interact with an AI and it's across 00:23:32.670 --> 00:23:39.049 a number of It's tied into the common European 00:23:39.049 --> 00:23:43.289 framework. And so the stated goal is to help 00:23:43.289 --> 00:23:47.390 students get from B1 to B2 level and then from 00:23:47.390 --> 00:23:53.150 B2 to C1 level. It's not to produce, it's somebody 00:23:53.150 --> 00:23:56.349 from the beginning and it's not to take them 00:23:56.349 --> 00:23:59.210 to the high end. So I like the fact that it's 00:23:59.210 --> 00:24:02.450 kind of scaled in that way with a progression. 00:24:02.940 --> 00:24:08.839 A second one, which I'd also seen at Calico last 00:24:08.839 --> 00:24:13.819 year, Pangea Chat. And this is run by Will Jordan 00:24:13.819 --> 00:24:19.200 Cooley. And it started mostly as an idea of texting 00:24:19.200 --> 00:24:24.180 or audio chat with friends. And those friends 00:24:24.180 --> 00:24:28.660 can be actual friends or they can be people in 00:24:28.660 --> 00:24:32.589 other places that are co -learners with you. 00:24:33.150 --> 00:24:35.630 And I believe the main thing there was the AI 00:24:35.630 --> 00:24:38.369 would support and giving feedback on the interaction. 00:24:39.230 --> 00:24:45.150 But now they've gone to having a full on AI chat 00:24:45.150 --> 00:24:47.589 bot where you can practice without another human 00:24:47.589 --> 00:24:52.390 just with the chat bot. It helps you produce 00:24:52.390 --> 00:24:56.029 your target language. It helps you understand 00:24:56.029 --> 00:25:01.160 what's being said to you and just Kind of helps 00:25:01.160 --> 00:25:04.519 you learn on the go. So that one I'm interested 00:25:04.519 --> 00:25:08.420 in learning a little bit more about one called 00:25:08.420 --> 00:25:12.480 Unicorn Tutor AI and the guy talked to there 00:25:12.480 --> 00:25:20.559 was Albert Kano. And this one is. It's more of 00:25:20.559 --> 00:25:24.380 a. It's almost like a learning management system. 00:25:25.099 --> 00:25:30.200 with AI built in where an institution would, 00:25:30.200 --> 00:25:33.980 for example, feed its curriculum in, they turn 00:25:33.980 --> 00:25:39.240 it into AI lessons. So they build off of the 00:25:39.240 --> 00:25:42.819 material so that you can have conversations with 00:25:42.819 --> 00:25:45.740 avatars about the material. And those avatars 00:25:45.740 --> 00:25:49.619 can be playing roles of characters within the 00:25:49.619 --> 00:25:53.500 material or playing more of a tutorial role. 00:25:53.789 --> 00:25:57.329 I didn't get very deeply into it, but it was 00:25:57.329 --> 00:26:03.670 interesting to me because it has very good tracking 00:26:03.670 --> 00:26:07.289 for teachers. Although like a lot of the tracking, 00:26:07.529 --> 00:26:11.809 I think it provides so much information for teachers 00:26:11.809 --> 00:26:15.410 to actually go through and see what's going on. 00:26:15.730 --> 00:26:17.769 My guess is what they're going to, and maybe 00:26:17.769 --> 00:26:21.289 they already do this, is build an agentic AI 00:26:21.289 --> 00:26:24.549 on top of it that actually does a lot of the 00:26:24.549 --> 00:26:27.630 tracking, interpreting of the tracking data, 00:26:27.630 --> 00:26:30.329 and then makes recommendations for the teacher 00:26:30.329 --> 00:26:32.869 to look, you know, into something in more depth. 00:26:33.289 --> 00:26:37.430 I don't know yet. Albert, if you're listening, 00:26:37.710 --> 00:26:40.390 maybe you can contact me and tell me what I've 00:26:40.390 --> 00:26:43.289 got right and wrong here. It's also white label, 00:26:43.690 --> 00:26:46.609 which is a term I'd heard before, but didn't 00:26:46.609 --> 00:26:49.029 remember exactly what it meant. And that means 00:26:49.029 --> 00:26:54.180 you give them your stuff and you put what they 00:26:54.180 --> 00:26:56.759 deliver to the students and to the teachers actually 00:26:56.759 --> 00:27:01.019 has your institutional name on it. So they, rather 00:27:01.019 --> 00:27:04.660 than presenting themselves as Unicorn Tutor AI. 00:27:05.420 --> 00:27:10.299 So that's interesting. The last one, something 00:27:10.299 --> 00:27:14.519 called Chattybots, and this is from eslvideo 00:27:14.519 --> 00:27:19.339 .com, where students take existing materials 00:27:19.339 --> 00:27:23.880 that they have So these can be conversations, 00:27:24.140 --> 00:27:27.539 role plays. There's one, just describe a picture 00:27:27.539 --> 00:27:31.220 so they provide the picture. And then you interact 00:27:31.220 --> 00:27:34.019 with the AI conversationally to do whatever it 00:27:34.019 --> 00:27:37.019 is. The role play, of course, can be a role play 00:27:37.019 --> 00:27:40.619 and describing the picture. Presumably the AI 00:27:40.619 --> 00:27:43.220 knows something about the picture too and then 00:27:43.220 --> 00:27:45.920 can tell you what you're getting right and wrong 00:27:45.920 --> 00:27:49.410 and so on. This also has the option of teachers 00:27:49.410 --> 00:27:51.569 building their own chats to go with the pictures 00:27:51.569 --> 00:27:58.569 and videos. Okay, the last one is the publications. 00:27:59.190 --> 00:28:03.829 And TESOL, I think if I got it correct, has only 00:28:03.829 --> 00:28:07.329 three publications out at the moment that are 00:28:07.329 --> 00:28:11.740 sort of teacher materials in AI. Other things 00:28:11.740 --> 00:28:14.759 they have I suspect have more and more AI built 00:28:14.759 --> 00:28:17.960 into them So one of these is what they call a 00:28:17.960 --> 00:28:20.200 zip guide and they have this for various topics 00:28:20.200 --> 00:28:27.140 But one of them is AI in ESL EFL It's six pages 00:28:27.140 --> 00:28:32.279 on plasticized paper so that you can carry it 00:28:32.279 --> 00:28:36.000 around and Not worry about it getting damaged 00:28:36.000 --> 00:28:39.859 has a lot of information distilled into those 00:28:39.880 --> 00:28:46.019 six pages, has sections on increasing AI expertise, 00:28:46.599 --> 00:28:52.480 using AI for specific needs. One page is focused 00:28:52.480 --> 00:28:57.200 on prioritizing process over product, which is 00:28:57.200 --> 00:29:01.319 something I like. It's where students, for example, 00:29:01.359 --> 00:29:07.339 learn about AI limitations, ethics, and creating 00:29:07.339 --> 00:29:10.460 guidelines for student use. One thing I found 00:29:10.460 --> 00:29:13.980 interesting in the introduction, there's a list 00:29:13.980 --> 00:29:19.259 of statements about AI. The first one is something 00:29:19.259 --> 00:29:25.779 along the lines of, it's okay not to use AI if 00:29:25.779 --> 00:29:29.559 you're uncomfortable with it. That's, in fact, 00:29:29.839 --> 00:29:36.619 number one in their list of statements. Six pages 00:29:36.619 --> 00:29:42.200 already, right? Yeah, so that's, I mean, we've 00:29:42.200 --> 00:29:46.220 talked about this. It comes down to, yeah, you 00:29:46.220 --> 00:29:48.619 don't have to use AI, but you need to know about 00:29:48.619 --> 00:29:55.440 AI. Right. Because, well, we believe that anyway. 00:29:56.759 --> 00:30:01.230 So the second one is a book. called AI Enhanced 00:30:01.230 --> 00:30:04.309 ELT, Innovative Strategies to Transform Your 00:30:04.309 --> 00:30:07.690 Classroom. That one on the inside cover, it has 00:30:07.690 --> 00:30:12.089 nine best practices, covers a lot of what we 00:30:12.089 --> 00:30:20.250 do in our framework, ethics, prompting, assessment. 00:30:21.190 --> 00:30:25.009 What I really found useful in it for me is it 00:30:25.009 --> 00:30:30.390 has a large number of tools. that some of which 00:30:30.390 --> 00:30:33.349 are language specific, some of which are language 00:30:33.349 --> 00:30:35.490 learning specific, some of which are general, 00:30:35.910 --> 00:30:38.210 but how they can be adapted for language learning. 00:30:39.089 --> 00:30:42.269 And yeah, the most comprehensive list I've seen 00:30:42.269 --> 00:30:46.869 to date with annotations about them presented 00:30:46.869 --> 00:30:51.289 both as short paragraphs about each one and also 00:30:51.289 --> 00:30:56.769 in a nice table so that you can see which ones 00:30:56.769 --> 00:30:59.980 do what sorts of things. So that's something 00:30:59.980 --> 00:31:03.279 I want to get into a little more Deeply because 00:31:03.279 --> 00:31:06.319 as you know since I'm I'm not currently teaching 00:31:06.319 --> 00:31:11.119 a language class I I don't pay as much attention 00:31:11.119 --> 00:31:14.779 to this as I do the more general generated eyes 00:31:14.779 --> 00:31:22.799 and the third one is It's a box consisting of 00:31:22.799 --> 00:31:27.819 a hundred no prep activity cards for AI and these 00:31:27.819 --> 00:31:32.940 activity cards just have anything from five minutes 00:31:32.940 --> 00:31:38.460 to about 30 minute activities. You pick one out 00:31:38.460 --> 00:31:42.400 at random and it says, here's what the goal is. 00:31:42.420 --> 00:31:46.380 They have them divided by skill area and so on. 00:31:47.799 --> 00:31:51.819 For example, number 60 is called dramatic storytelling. 00:31:53.000 --> 00:31:56.549 It says, open an AI in voice mode. and then gives 00:31:56.549 --> 00:32:01.150 you a prompt, tell me a story about X. And then 00:32:01.150 --> 00:32:06.470 now tell it again in a different voice. So an 00:32:06.470 --> 00:32:09.730 angry voice, an excited voice, a sad voice. And 00:32:09.730 --> 00:32:13.710 then you have students pair up and discuss what 00:32:13.710 --> 00:32:17.470 they notice different. The goal is to get them 00:32:17.470 --> 00:32:22.670 a little more aware of how voice tone and intonation 00:32:22.670 --> 00:32:26.730 impact the meaning of something. So I think it 00:32:26.730 --> 00:32:29.670 might need a better prompt, but the fact that 00:32:29.670 --> 00:32:31.549 these things are out there, they're a good place 00:32:31.549 --> 00:32:34.470 to start because they are specific to language 00:32:34.470 --> 00:32:38.569 teaching and they've been done by people who 00:32:38.569 --> 00:32:41.130 know something about language teaching rather 00:32:41.130 --> 00:32:45.150 than by people who know technology and not something 00:32:45.150 --> 00:32:47.890 about language teaching. So I hope to see more 00:32:47.890 --> 00:32:52.829 of this in 2027. This is especially the last 00:32:52.829 --> 00:32:56.720 part. concrete materials that TESOL seems to 00:32:56.720 --> 00:33:00.240 put out there. It sounds to me like this is also 00:33:00.240 --> 00:33:03.700 relevant for other language teachers or teachers 00:33:03.700 --> 00:33:09.240 of other languages other than English. In general, 00:33:09.460 --> 00:33:12.599 coming back to the actual conference experience 00:33:12.599 --> 00:33:18.819 and participating in a conference, what can teachers 00:33:18.819 --> 00:33:22.720 do? What are the next steps? You started talking 00:33:22.720 --> 00:33:25.160 off about collaboration. How does this actually 00:33:25.160 --> 00:33:29.160 work? Yeah. Well, I mentioned my own take on 00:33:29.160 --> 00:33:33.279 this as we've gone through. I do think it's important 00:33:33.279 --> 00:33:38.519 as I did to personalize it. So I picked on things 00:33:38.519 --> 00:33:41.799 that I thought were either interesting to me 00:33:41.799 --> 00:33:45.559 intrinsically or interesting to me in trying 00:33:45.559 --> 00:33:53.430 to build my value. as a teacher educator. So 00:33:53.430 --> 00:33:57.849 what do you think? I agree. It has a lot to do 00:33:57.849 --> 00:34:00.650 with the people. And I think there's even better 00:34:00.650 --> 00:34:03.069 saying, especially in the times of artificial 00:34:03.069 --> 00:34:05.549 intelligence, it's about the people, basically. 00:34:05.970 --> 00:34:08.630 Of course, one listens to these talks or takes 00:34:08.630 --> 00:34:10.929 photographs of the slides, takes notes and thinks 00:34:10.929 --> 00:34:16.889 that that's all very good. But I think I always 00:34:16.889 --> 00:34:21.469 get the most out of Often these conversations 00:34:21.469 --> 00:34:25.070 happen during breaks or in the evenings, actually 00:34:25.070 --> 00:34:29.429 talking to people. Because once you know a little 00:34:29.429 --> 00:34:33.510 bit more about the person, it becomes embodied 00:34:33.510 --> 00:34:38.829 knowledge almost basically. Things get clearer 00:34:38.829 --> 00:34:42.710 for me in some way if I have the chance to interact 00:34:42.710 --> 00:34:47.190 with the presenter, with an author of some... 00:34:47.130 --> 00:34:52.829 tool or material or whichever one it is. Well, 00:34:52.849 --> 00:34:56.769 you've never been to TESOL, but there are lots 00:34:56.769 --> 00:35:01.329 of lots of other conferences. Maybe we should 00:35:01.329 --> 00:35:04.369 do the same thing since we're both going to Calico 00:35:04.369 --> 00:35:08.659 and Uricall this year. As I always say, my English 00:35:08.659 --> 00:35:11.599 is not good enough for TESOL, basically. That's 00:35:11.599 --> 00:35:15.260 why I didn't go. Of course, teaching German, 00:35:15.519 --> 00:35:18.900 there's kind of less reason for going to TESOL. 00:35:19.239 --> 00:35:22.719 But yeah, certainly those conferences on computer 00:35:22.719 --> 00:35:25.800 -assisted language learning that we both often 00:35:25.800 --> 00:35:29.760 attend are a good opportunity. Let's think of 00:35:29.760 --> 00:35:33.699 something. I'd like to get our slogan in here, 00:35:33.739 --> 00:35:37.840 because this also, I think, applies to conferences 00:35:37.840 --> 00:35:41.280 in some way. Let's remember, artificial intelligence 00:35:41.280 --> 00:35:45.400 is no substitute for natural ignorance. Well, 00:35:45.420 --> 00:35:49.280 indeed. Let's keep learning. Till next time. 00:35:57.480 --> 00:36:00.519 We are grateful for the support for opening AI 00:36:00.519 --> 00:36:03.199 for language learning by the Language and Applied 00:36:03.199 --> 00:36:05.679 Research Center at San Diego State University 00:36:05.679 --> 00:36:08.159 and the Southern Area International Languages 00:36:08.159 --> 00:36:11.179 Network, SAILN, which is part of the California 00:36:11.179 --> 00:36:14.280 World Languages Project. Mari Ocando Finol is 00:36:14.280 --> 00:36:17.019 the production coordinator of OAILL. Our theme 00:36:17.019 --> 00:36:20.639 music is by Tillmann Spiegl. Our editor is Chris 00:36:20.639 --> 00:36:23.820 Brown. Live conversations are moderated and promoted 00:36:23.820 --> 00:36:26.719 by me, Shahnaz Ahmadeian. Until next time.