Bonni Stachowiak [00:00:00]:
Today episode number 521 of the Teaching in Higher Ed podcast, the myth of the AI first draft with Leon Furze. Production Credit: Produced by Innovate Learning, maximizing human potential. Welcome to this episode of Teaching in Higher Ed. I'm Bonni Stachowiak, and this is the space where we explore the art and science of being more effective at facilitating learning. We also share ways to improve our productivity approaches, so we can have more peace in our lives and be even more present for our students. I'm so pleased today to be welcoming to the show Leon Furze. He's an international consultant, author, and speaker with over 15 years experience in secondary and tertiary education and leadership. Leon is studying his PhD in the implications of generative artificial intelligence on writing instruction and education.

Bonni Stachowiak [00:01:12]:
Leon has held roles at multiple levels of school and board leadership, including director of teaching and learning, head of English and elearning. Leon is a nonexecutive director on the board of Young Change Agents and Reframing Autism, and a member of Council for the Victorian Association For the Teaching of English. Leon completed his master of education at the University of Melbourne in 2016 with a focus on student well-being, leading schools through change, and linking education systems and community. He's published dozens of books, articles, and courses with his most recent publications, practical AI strategies, practical reading strategies, and practical writing strategies reaching an international audience. Leon presents at state and national conferences and runs online and face to face professional learning for schools, individuals, and businesses. Through consultancy and advisory work, Leon helps educators from k through 12 to tertiary to understand the implications of general artificial intelligence in education. Leon Furze, welcome to Teaching in Higher Ed.

Leon Furze [00:02:31]:
Thank you very much.

Bonni Stachowiak [00:02:33]:
Would you start by telling us what your writing practice looked like, oh, let's say, 5 years or more ago? What did it used to look like?

Leon Furze [00:02:41]:
Wow. Yeah. I've I've been a writer for quite a long time, I guess. You know, I was a writer when I was young. I used to write a lot, but then when I became a a teacher, a secondary teacher, I started writing around about 2016, 2017 for that audience. So my colleagues, the the State English Teachers Association. So the writing then was was very, very geared around professional practice and exploring ideas around pedagogy. A couple of years after that, I started to dabble with fiction.

Leon Furze [00:03:15]:
So I was I've written quite a bit of, short science fiction, speculative fiction, and I've always enjoyed doing that. And then over time, expanded out to to writing longer books. All the books are nonfiction. So, again, moving back to that kind of teacher professional practice and and pedagogy areas. I haven't managed to pull off a full length fiction book yet. Although I think I would love to do that one day, but I'll stick to short fiction for now.

Bonni Stachowiak [00:03:42]:
And as you think about writing as someone who writes yourself, as someone who teaches others to write, what are some of your concerns about the value that education tends to place on the ability to write?

Leon Furze [00:03:55]:
Yeah. It as a as a former English teacher and a writer myself, it does pain me a little to say this, but I I think that we put writing on a pedestal in education, both in k to 12 and higher education. It disadvantages a lot of students for whom writing isn't natural. It's not something that they're confident with. And it really means that we we don't capture those students, their knowledge, their skills. We prioritize writing a lot. We treat it as the ultimate endpoints. And I think that disadvantages

Bonni Stachowiak [00:04:33]:
And now as if that wasn't a complicated enough thing for us to start with, I'm gonna just just pile one more, easy topic on there and and layer Bonni. Could you share a bit on how AI has started to change these conversations about the value of writing?

Leon Furze [00:04:51]:
Yeah. I started my PhD in in generative AI and writing instruction in November 2022. So 2 weeks before ChatGPT was launched was the official start date of my

Bonni Stachowiak [00:05:02]:
PhD. Incredible.

Leon Furze [00:05:04]:
My supervisor and I were talking. We were saying, you know, these technologies, they're out there, GPT 2, GPT 3, and they're going to come into the classroom. They're gonna really impact, education and particularly writing instruction over the next 3 to 5 years. And then 2 weeks later, ChatGPT came out and we said, hold on. I think we need to shorten this timeline. It's probably gonna be months, not years. And we've seen that since November 22, huge kind of narrative in the public sphere around the threat to writing, the threat to education more broadly, English and essay writing, really specifically. And I think that a lot of that comes from that idea that writing is just an assessment practice.

Leon Furze [00:05:45]:
So the biggest threats that I've seen from AI in in regards to writing is that it it challenges those those notions that students need to be able to write well in order to express their knowledge because now we have a machine that can write as well as, if not better than, a lot of people. I think that that that really has given us in in education, both in k to 12 and tertiary, perhaps an invitation to change assessment practices, to shift them away from being primarily writing in focus, and also to rethink why we actually teach writing.

Bonni Stachowiak [00:06:21]:
And you have come up with this idea of the myth of the AI first draft in terms of some of the concerns that have started to emerge as the technology continues to emerge. Could you describe this myth?

Leon Furze [00:06:35]:
Yeah. Sure. It's, really from my observations in how, educators were talking about these technologies and how EdTech and technology developers were pushing these technologies, A lot of the initial use cases I saw for generative AI were talking about getting students to create their first drafts with AI. I even saw this coming from the big developers like Microsoft. They put out a video around accessibility and using Copilot if you have disabilities. And this narrative of creating a first draft with AI is problematic for me because it devalues what I think a first draft is for. I think the first draft really is the opportunity to just dump ideas down on the page. It's an opportunity to clarify your own ideas, to synthesize things that you've pulled in from other places.

Leon Furze [00:07:28]:
It's an opportunity to create knowledge, and just offloading that section of the writing process to AI, I think, minimizes the importance of the draft. The other thing is it it kind of it shifts the focus from the purpose of writing again towards assessment as an outcome or an endpoint because it says, okay. The important part is not the first draft. The important part is the thing that you create from the first draft. So let AI do that unimportant first draft work, and then you do the really important bit of actually submitting an assessment piece, which I think is totally backwards.

Bonni Stachowiak [00:08:03]:
And talk a bit more then if if the purpose of writing isn't to assess one's learning, then what should be or what could be the purpose of writing?

Leon Furze [00:08:13]:
I I've always seen writing as, a creative act rather than an act of, I don't know, performance or repeating things that you've learned in the past. You know, if I'm if I'm reading something, if I'm studying, for example, like, with my PhD studies, I'll take a lot of notes. I'll synthesize a lot of ideas from other people. But when I write, it's really to solidify my own knowledge and to then create something new and interesting. I think that kind of creative aspect of writing is true whether it's analytical writing, persuasive, creative, or imaginative writing. So for me, the purpose of writing is not just to show knowledge, that you've gained along the way. It's actually to turn that knowledge into something new, whatever form that ends up being.

Bonni Stachowiak [00:09:02]:
It's interesting as you're as you're sharing because I'm thinking about just the writing I've been doing in the last 48 hours. I belong to this group of people who it's a it's a special interest group, part of a professional association that I belong to, and someone sent out information about Google Notebook l m, and I was experimenting with it. And for those listening who may not be familiar with it, you can upload rather large text files to it. You can also upload PDFs. And rather than go out to a large language model like, chat GPT or Claude or or one of those, it's basically like creating your little small I mean, relatively speaking, a a small language model. So I got very to get to get me to spend a lot more time on this experimentation than I normally would Dave. But I I uploaded all of the artificial intelligence episodes, transcripts that I've done on teaching in higher ed, as well as some of the milestone episodes, episode 102 100, etcetera, and also ones about assessment and grading. And I started asking that questions, and I'm I'm thinking as you're the purpose of writing that probably what I was doing there was not necessarily a creative act, but it was a result of taking all these creative acts that Dave already done and then seeing what this smaller language model might do with that.

Bonni Stachowiak [00:10:39]:
So I think maybe perhaps it's helpful for us to distinguish between synthesizing our prior creative acts or the creative acts of others versus that this idea of a blank slate. And that's something that was super intriguing to me as I looked at your and and read through your blog, and and also intriguing of the tension that there is between wanting to avoid for so many people, the pain and agony of the blank slate, how do you get started, the writer's block, and whether or not it's helpful to have that pain or agony be diminished in some way. So would you share a little bit about this tension between the blank slate and how important you see that in this creative act of writing and or or whether or not you you find that to be an important tension as as someone who writes and communicates.

Leon Furze [00:11:32]:
Yeah. Absolutely. I, I really like Margaret Atwood's writing about about the blank slate and and getting over the fear of the blank slate. I think the blank slate's a a great place to be at. You know, the blank page for me isn't something that that represents fear or causes any kind of anxiety, but I know that it is for a lot of people. But what what I feel is that it's we have to shift our perceptions around writing and around creativity Dave from this idea that the blank page is scary or the blank page is something to be overcome as quickly as possible and see it as an inherent and and very important part of that struggle. One thing I think we've done in education with writing where because we've devalued it, because we've reduced it to an assessment outcome, we've also created this idea that it's okay to take shortcuts. It's okay to just skip ahead to the end point.

Leon Furze [00:12:25]:
It's okay if you're not that good at at at doing it to, to kind of look for little hacks and cheats and things. Realistically, I think that writing is hard. It's I mean, this this sounds kind of a little bit maybe banal to say so, but writing is a difficult thing, and that's okay. You know? Creating art is difficult. Creating software is difficult. Creating anything is a difficult process. So when we devalue writing and we say, okay. There are shortcut shortcuts or you can get over the blank page with this AI generator or anything like that, We're kind of saying writing doesn't have to be hard.

Leon Furze [00:13:01]:
I I would say, look. Let let writing be something which is difficult. Let writing be something which some people gravitate towards and others don't. And the people that don't enjoy writing or the people that for for whatever reason don't engage with that form, Let them do whatever works for them. Let them express themselves in whatever ways they feel comfortable or confident with or in ways which they can really lean into their own kind of creativity. Everyone is creative in some way, and everybody has the capacity to create, but we don't have to pretend that that's an an easy thing.

Bonni Stachowiak [00:13:34]:
And how would you then think about this is if this is something that is hard to do, the importance of doing it anyway, you know, and and discovering that. I I think often of this book that I read called Range, which was about sort of the propensity, especially here in the United States, where we will have our children hyper specialize, and they give lots of examples from things like sports or things like chess and, the benefits of sort of interdisciplinary approaches to learning, which is why I don't get worried when one of our kids decides that, this week he's really into coding, and if next week it's something else, I'm not gonna be so worried about the the changes. But so the idea of trying something and it's hard, and then giving up versus trying something, and it just really, another way would be better for us to express our learning. I guess what I'm trying to ask is to what degree is it that we might encourage others to do things even though they are hard versus what comes when we open up the freedom to to demonstrate our learning uses all using alternative approaches.

Leon Furze [00:14:48]:
Yeah. First of all, that's a great book. That's

Bonni Stachowiak [00:14:50]:
Oh, have you read it?

Leon Furze [00:14:52]:
David Epstein's book.

Bonni Stachowiak [00:14:53]:
and range. Yeah. It's

Leon Furze [00:14:55]:
I really like that idea of the importance of being a generalist, someone who writes a lot about AI. He he posts a lot of things on LinkedIn. Conor Grennan posted something recently about how well placed generalists are in this kind of technological era because AI is a very generalist kind of technology. One thing I think is that we can encourage people to do hard things. We can encourage people to experiment. I think, fundamentally, that should be part of k to 12 education to give students the the the space and the time to experiment, to find what areas they gravitate towards, to find what areas they need to push out a little bit more. We need to provide opportunities for people to find things hard, to persevere, and to see if they come out the other side. And, you know, some some people will find some things hard and eventually, they'll come to the conclusion that it's not for them.

Leon Furze [00:15:49]:
I I can't kick a ball to save my life. I grew up in England, very big on football, soccer. I was I spent countless hours out on the field trying to kick a ball in a straight line and could never achieve it. Moved to Australia where Australian football is very popular. Different shaped ball. Still can't kick it.

Bonni Stachowiak [00:16:06]:
Stuck you out.

Leon Furze [00:16:07]:
And now I've decided that kicking a ball is just not for me. But there are other skills that I've worked at over time where I have seen improvements, and that's, you know, given me that little boost. It's given me the encouragement, and I've been able to lean into those skills. So why we why we seem to have this assumption in in education that, first of all, everybody needs to be equally good at some things, and then to really quickly specialize into very niche areas, it's it's beyond me. I'm not sure why why we do that.

Bonni Stachowiak [00:16:37]:
Yeah. I'm thinking about the as I reflect on my own successes, I put that in air quotes, but and then, certainly, a lot of challenges as AI continues to emerge and be more universally adopted. I I think about the people that I've talked to on the show before about writing and the ways in which we privilege certain types of writing, and that really negates our ability to even know whether or not somebody actually understood something. I I've always thought often thought about the basic communication process between, you know, a message gets sent, and it was packaged up in a in a particular context, and then I need to, as the receiver, interpret that and and then be able to synthesize what I think about it. And when we're so specific about, oh, it wasn't in the right format, and you didn't write with the right grammar. You know, we're we're, losing opportunities to see what happened in that process. And I it seems to me that artificial intelligence has only amplified and magnified the possibilities for people's unique and distinct voices to be lost in their own learning process, and it's really sad and and quite discouraging too. When when I see it happening, I do try to remind myself that this is not something being done to me.

Bonni Stachowiak [00:17:55]:
This is a result of a lot of other factors going on both in this individual's life, if it's happening with the student or also just in our educational systems more broadly speaking. I guess this comes back to you say there are some ways that we can do better and change the definition of what better looks like. Could you talk some about better?

Leon Furze [00:18:17]:
Yeah. I mean, it's, again, it's it's a a bit of a paradox in education that we we encourage students to excel, but the way that we do that is by pushing standardization. Mhmm. So we say, you all have to have a certain level of expertise in in literacy or numeracy, for example, And that level is here, and we kind of set this arbitrary bar. But then to encourage people to to be to be better at certain subjects, we don't spend any kind of time or dedicate any energy to actually allowing students to do that. We get so bogged down in the in the standardization that we never actually allow students to get Bonni to the the really good parts. I mean, one one thing I found the more I've gone through adult education, you know, I did my masters in in 2016. I'm doing my PhD now.

Leon Furze [00:19:07]:
It's a totally different model of learning to what we subject students to in k to 12 education. There is no handholding. There's no there's no sort of standard bar to achieve. It's just do as well as you can with the materials that you have at your disposal. And so I think in in education, to encourage people to be better, we have to give them the time and the space to to explore and to experiment.

Bonni Stachowiak [00:19:33]:
And would you talk specifically you've talked about the myth of the AI first draft. Would you share some of the alternatives to the AI first draft? What are some of these ways that we might go about thinking differently?

Leon Furze [00:19:46]:
Yeah. Sure. So I I think that I mean, the first thing that should come before any kind of piece of writing is some kind of discussion or some kind of idea generation process. I wrote another piece recently about brainstorming, which is another one of those really kind of easy low hanging fruit uses of AI. People are suggesting just use AI to do your brainstorming. And sometimes that's okay, but a lot of the time that just kind of supplants the creative process. Again, for me, brainstorming, idea generation, drafting, all of that is is far more important than the finished product itself. So I would say as an alternative to getting AI to do the first draft, you might do some brainstorming yourself or you might work with your peers or or in a in a a community of writers or just in a classroom context and discuss and capture some of that discussion.

Leon Furze [00:20:39]:
So one thing I talk about in the article is maybe capturing voice transcriptions of those discussions and then using AI to sort of tidy up those notes. And then once you've got your organized notes, your your your brainstorm, the idea dump, or whatever, just have a crack at writing some sort of first draft yourself before then perhaps ending handing it off to, AI to tighten up some of the language. I think that the thing that AI is really good at is the kind of functional spelling, punctuation, grammar, all of that end of the writing process. So, you know, let AI take care of that stuff, but reserve the idea generation, the creativity, the synthesis of ideas. That can be something that we still do. One sort of counterpoint that that I've had when I've published that article is that it might be seen as a bit of a an elitist approach to writing to say, you know, writing is an inherently creative act and it should be difficult. It should be a struggle. It can sound a little bit elitist.

Leon Furze [00:21:40]:
It can sound like it excludes people who, for whatever reason, can't or won't write. To which I would argue, I have never met a a young person, a student who is incapable of creating something. I've never met a person who's incapable of creating something. And so whatever way they are capable of creating, we should be supporting that and not trying to kind of pigeonhole them into the writing mode. So, you know, if students are able to create things through art or verbally or through discussion or in any way, really, capture that creativity. And then if necessary, use AI to help turn that into a piece of formal writing at the end of the process.

Bonni Stachowiak [00:22:22]:
I imagine that we spoke about the book range and just the benefits of interdisciplinary thinking. Imagine if that were possible to do more of, and it really does take a risk, not just at the individual level, but also at the collective level too. Would you would you speak a little bit more about what these alternatives might look like inside of an individual class? What what how how would we create the trust that is necessary for somebody who doesn't have a practice of writing to consider brainstorming without jumping straight to what may seem like the solution of AI?

Leon Furze [00:23:02]:
Trust is is probably a a good word to use here, I think. I think we've we've broken a lot of trust in in education when it comes to assessments, writing, high stakes testing, and all of those areas because we Dave, in effect, said the point of assessment is to achieve this outcome. And, as I said, with standardization, we've set an arbitrary bar and said, you you must be this literate in order to pass this course. You must be x, y, and zed. And that that doesn't create an environment of trust in the classroom. So I've I've been writing a little bit recently around other approaches to assessment, so things like ungrading, removing that final grade from an assessment task, removing the the endpoint essentially, authentic assessments and performance based assessments or programmatic assessments that run throughout an entire course. And I think all of these kinds of alternative methods of assessment, they shift the focus away from the endpoints and hopefully encourage students to see that it's the it's the the journey of learning is the more important part than just the outcome at the end. And like I said earlier, some students will gravitate towards writing.

Leon Furze [00:24:16]:
Some students will enjoy writing and will do well and will maybe struggle through and become, excellent writers. That's not for everyone, but if we can shift that narrative around the endpoint, around the assessment goal, then maybe we can allow people to take other journeys to get to the same point.

Bonni Stachowiak [00:24:36]:
Before we get to the recommendations segment, I just wanted to ask you a question about the other side of this equation. We've talked a lot about the writer, the the creator, the thinker, the creative person. What about the person who is either assessing their work or perhaps even in an ungrading model just facilitating some parts of that process. I'm very concerned about the ways in which we talk about students' use of AI as immediately will go to cheating. But when professors, faculty members would use AI for grading purposes. That doesn't seem to be met with the same idea that the word cheating tends to bring up for us. So any reflections for us on the use of AI on the other side of the equation for assessing work?

Leon Furze [00:25:24]:
Yeah. Probably more than we've got time for.

Bonni Stachowiak [00:25:27]:
I was imagining that as

Leon Furze [00:25:28]:
I This is a very contentious issue, isn't it, at the moment? Look, we're we're in the same position as as as you are in the States. So here in Australia, we have a teacher shortage. We have huge issues with workload and burnouts. We have issues in just getting kids into the classroom as well as finding teachers to put in the classrooms. So you can see how this is a perfect storm for this technology to be seen as a way through and a way out, perhaps. And there are already working parties across Australia that are exploring things like teacher workload, can AI be used for assessment purposes, and so on. I mean, I sort of reading between the lines of of of your question. If we're going to accuse students of cheating and then allow professors or or educators to use the technology for assessments, that's hypocritical, and probably quite condescending as well because it says, okay.

Leon Furze [00:26:19]:
We're the adults. We can use the technology responsibly. You're the kids. You can't. And I have some real problems with that. I I also have some problems with it from a tech a technological standpoint. So I think that these technologies, as they exist now, are fundamentally incapable of making qualitative values based judgments about text. They are predictive text models.

Leon Furze [00:26:46]:
They might look like they're capable of thought. They might look like they're capable of reason, but they're not. They're just sophisticated prediction models. And that means if we give a student's work to an AI and say, grade this or assess this piece of writing against this rubric, it will do a really convincing job of looking like it's grading, but it's not grading. It's not assessing. It might do an okay job of spelling, punctuation, grammar, maybe even structure, cohesion, logic, some of those other functional aspects, but it can't make an aesthetic judgment. It can't tell you if a piece of student's creative writing is beautiful or worthy or any of those kinds of subjective things that we might be looking for in a student's writing. It can tell us if an analytical piece of writing is clearly argued, but it can't tell us if that student is engaged with the material on a personal level because it doesn't know what a person is.

Leon Furze [00:27:45]:
It doesn't know anything. In fact, it's just predictive text. So that's for me, that's a fundamental technological barrier to using the technology for assessment purposes, which is not being addressed because it's expedient and because technology developers certainly would have us believe that this technology is gonna solve all of those problems. And the other side of it is this is students' intellectual property. I have huge problems with software that already exists where educators have been encouraged to just dump students' intellectual property in for the purposes of things like plagiarism checking, and these companies are now worth 1,000,000,000 of dollars. So, you know, this is just another way for technology companies to profit from student intellectual property, which, yeah, I'm not particularly enthusiastic about. I'm not sure if that came across. I think there there are reasons that we might want to investigate AI for assessment purposes, workload included, but we can look at students using it for self assessment.

Leon Furze [00:28:48]:
We can look at students using artificial intelligence to help improve their own writing before submission. Ultimately, a human needs to discuss the value, the aesthetics, and some of those subjective qualities of the writing because an AI language model can't do that.

Bonni Stachowiak [00:29:06]:
Well, for someone who said that was a whole another conversation, you sure did get a lot in there in a very concise way. Thank you so much for those reflections and for bringing so many of those things to light.

Leon Furze [00:29:17]:
You're welcome. This is

Bonni Stachowiak [00:29:18]:
the time in the show where we each get to share our recommendations, and you were almost talking about about mine in that not that not by its brand name, but I've really grown accustomed to using this new app called Whisper Memo. I will tell people, as a cautionary note, there are a lot of apps out there that talk about something called Whisper. Whisper is just a way of taking audio to when you talk, making that into text. But so this is you wanna specifically look for whisper memo is the one that I'm talking about in the App Store. And why I like it is that both on my Apple Watch as well as on my phone, I can just press a button or tap on a complication and begin speaking. It has, as of this recording, a 15 minute limit, which I had never just speak extemporaneously like that more than 15 minutes, so that hasn't been much of a limit. And it's using artificial intelligence to essentially create a transcript of what I'm saying. What makes it distinct from other similar services is that it will very intelligently put in the paragraph breaks and makes it that much easier when I go back to look at what I was saying.

Bonni Stachowiak [00:30:28]:
It just makes sense of it and and shortens that process a little bit for me to go back to it. They also just came out with this feature where if I say something like, remind me to do this, it'll put a little notification on my phone on the on the lock screen until I check it off. So it's got, like, kind of a neat little just capturing ideas and thoughts and brainstorms that I have throughout the day. If it's a longer thing, then it'll put that into paragraphs, and then you can set it up to email yourself, or I I actually have a special email address for another app that I use called Drafts. And so all of my thoughts go into the Drafts app, and they're there waiting for me. And it's just a wonderful app, very easy to get started, a relatively inexpensive price point, and I cannot recommend it enough if you like to capture your ideas as you're creatively thinking about things throughout your day. So, Leon, I'm gonna pass it over to you for whatever you'd like to recommend.

Leon Furze [00:31:24]:
Yeah. Sure. So my recommendations are always just centered on AI because that's that's all I all I think about at the moment with the PhD and with my consulting work. And just recently, I finished a book by an author, Marek Kowalkevich, and I've, probably butchered his name, which I apologize, Marek. We have I've actually spoken to Marek in the past as well on a Zoom call, so I should be able to pronounce his name. But his book is called The Economy of Algorithms. It's a a fairly new book. It's not just about AI.

Leon Furze [00:31:58]:
It's about algorithms more broadly and their their impact on society. But I enjoyed it because I like to look at technologies from different perspectives, and Marek's an economist, so well out of my wheelhouse. And and the book just gives a really nice kind of first of all, an introduction to what algorithms are, a bit of history of algorithms, what it means for social media technologies, generative AI, and and AI more broadly and a whole range of other technologies. But then he also starts to explore the impact that algorithms have on day to day life, the ways that businesses might approach algorithms, the way that governments and sort of entire societies might approach algorithms. And it goes from a a beginner's level through to a quite advanced academic level. Marek's a professor in a university over here in Australia. It never gets really bogged down in that kind of academic theory. It's very approachable all the way through.

Bonni Stachowiak [00:32:56]:
It also looks like it has the best subtitle of any book I've seen in a very long time, AI and the rise of the digital minions. That's just perfect.

Leon Furze [00:33:07]:
The digital minions. And and that's that's an idea I think he started playing around with a few years ago and and has just kind of expanded on this. But the the idea of treating these technologies as digital minions, not anthropomorphizing them, not pretending that they're some kind of revolutionary force or some force of nature, but just treating them as what they are. So, you know, I find that a really nice approach to talking about algorithms.

Bonni Stachowiak [00:33:29]:
Alright. Did you have anything else you wanted to recommend?

Leon Furze [00:33:32]:
I wasn't thinking of this, but I'm just gonna segue on from your whisper memo and just say that the the one I use, the version of the software that I use for that is called otter.ai.

Bonni Stachowiak [00:33:41]:
Mhmm.

Leon Furze [00:33:42]:
And like you said about WhisperMemo, it's a really capable transcription software. I've got a license, so it's it is unlimited, and I ramble on and on and on for hours in voice memos. So I record voice memos if I'm out for a long run or if I'm driving somewhere in Australia, do a lot of driving, maybe 4 or 5 hours of driving to get to somewhere where I'm working. So I record hours of voice transcription, and Otter catches it really well. It provides the little summaries, kind of does the structuring, and all of that kind of things. And I often take the transcripts from Otter into Claude 3, which is Anthropic's most powerful AI language model because Claude is really good at at parsing language much better than chat gpt. Sorry. That was 2 recommendations in 1.

Bonni Stachowiak [00:34:29]:
No. I loved it. This is so good. You are the have all the AI episodes that I have done, and I think I'm nearing 20 now. You're the first person to even speak about Otter I AI, which really surprises me because it I know that is rather prolific. It it when it started I don't wanna say it came out because I know it's been out for a while, but I think more people started using it. It must have a default feature or otherwise confuse people when they sign up because then they end up having their minions go attending all their meetings for them. And it created all this chaos at my university with, like, thinking people were doing this intentionally, and they were not necessarily always wanting their little minions to go attend every meeting they were scheduled to be at.

Bonni Stachowiak [00:35:13]:
It got, I was in a breakout room with 2 of these Otter.ai Bonni, and we didn't have a lot to say to each other. So I would Dave, actually, if you'd share even a little bit more about I mean, I guess you talked about the summaries and all that. But if you know anything about this calendaring function that if this has happened in any of your contacts, I'd love to hear about that.

Leon Furze [00:35:30]:
Okay. For exactly that reason. I'm I'm on a few I'm on a few boards. I've, you know, I've got 3 nonexecutive board roles. And in one of them, I was joining a board meeting for a not for profit young change agents, and my Otter AI joined alongside me. I was just audio calling in because I was driving, so I didn't see it join. And the board chair noticed, a few minutes into the meeting that this extra Leon was in the room and said, hey. What's going on? I think this meeting's being recorded, and we don't do that.

Leon Furze [00:36:00]:
Can we turn it off? And nobody could get rid of the the Otter AI. I had to pull over and go to a service station and, log out of the meeting, go to the Otter website, turn the feature off, turn off the auto join, and then log back into the meeting because nobody could physically kick it out. So I think that they've changed now and that that's it's off by default, and you have to opt into that. But it will join Google Meet, Zoom, Microsoft Teams, Skype. It'll join pretty much anything. It's very proactive, little AI chatbot. So, yeah, if you do use Octo AI, just make sure that it's not sort of ghosting into meetings on your behalf. It will actually also join meetings even if you don't.

Leon Furze [00:36:44]:
So, yeah, I've had meetings like you like you said where the person doesn't turn up or they're running late, but their their otter assistant is there sort of quietly sitting in the corner recording.

Bonni Stachowiak [00:36:55]:
And before we close out the episode, you talked about rambling on and on. Do I'm I was making some assumptions that I'm I'm remembering what it was like to do my doctoral research and how time intensive that is, but are there other things besides your doctoral research that you're rambling on and on about that may be helping other people spark a little creativity of how they might use something like Otter.ai?

Leon Furze [00:37:18]:
Yeah. Sure. I mean, I I use it for a lot of my article writing on my blog.

Bonni Stachowiak [00:37:24]:
Mhmm.

Leon Furze [00:37:24]:
So you mentioned the myth of the AI first draft. That was drafted on a on a run. I live on a farm just for context. I live on a farm about 35 kilometers from the nearest town in in regional Australia, so nobody can see me doing this, which is important. But if I'm out for a long run for a couple of hours and I've got my AirPods connected to my Apple Watch, I can record voice memos straight onto my Apple watch, and I often dictate entire draft articles.

Bonni Stachowiak [00:37:54]:
Wow.

Leon Furze [00:37:55]:
So, I mean, for me, this is just part of my writing process, and this this doesn't work for everyone. But I'm a very when I write, I I sort of visualize what the writing looks like as it's kind of unfolding in front of me. So as I'm out for a run, I can visualize paragraph by paragraph and just kind of verbally rattle off a lot of content. And then to take that recording and run it through, board with with a very specific prompt. So I've got the same prompt I use every time, which is something along the lines of, I will upload an audio transcript of a an article. You will retranscribe the article, correct any obvious transcription errors or repetitions, but, otherwise, do not change the language. Do not overwrite the text. If there is something which is very unclear in the recording, just put it in square brackets and leave it as is.

Leon Furze [00:38:48]:
And if I give explicit instructions to the language model during the transcription, follow those instructions. So that's a really specific prompt that I use.

Bonni Stachowiak [00:38:58]:
I I

Leon Furze [00:38:58]:
paste that into Ford, and then it says, okay. Yep. Sure. I'll do that. And then I paste in the transcript. Sometimes while I'm recording, while I'm out for the run, I might I might kind of say halfway through the the recording something like, this is an instruction for the language model. At this point, turn the following thing into a dot point list with bold headings and give formatting instructions, and it will it will do that as well. It will, you know, follow the instructions and format your text.

Leon Furze [00:39:24]:
So it's a bit weird. It certainly looks weird, I imagine, to anybody who drives past me while I'm doing this, but it works.

Bonni Stachowiak [00:39:30]:
And I I know that you have a book out about AI, and I know that you do you now and now I know a little bit more about how you are such a prolific blogger. Where would be some of the best ways for people who are curious like me? Where can I go as soon as we hang up and go start exploring and learning more from

Leon Furze [00:39:47]:
you? Yeah. Well, my website is just my Dave. Very uninteresting. It's it's just leonfurs.com, and the blog is there. I I I post blog posts maybe 2 or 3 times a week. So that's just leonferz.com/blog. The book is called Practical AI Strategies and was written for both k to 12 and tertiary educators. It's about using it yourself, not using it in the classroom with with students.

Leon Furze [00:40:13]:
And just this week, actually, I've just released an online course version of the book, which follows the structure of the book. So an introduction to AI, AI ethics, assessment practices, education policies, and how to actually use the technology. And, yeah, I've just turned it into a course, which is just over 4 hours long, broken into little lesson chunks. So the book and the course, all of the information for that is just at practicalai strategies.com. So, yeah, leonferz.com, practicalai strategies.com, and between those two things, you'll find pretty much everything I've ever written.

Bonni Stachowiak [00:40:47]:
Thank you so much. I'm so glad that we got to have this conversation, and I feel like it's kinda funny that I went first on the recommendations because now I just wanna go and experiment with what you're talking about and start to build my confidence there. I I find so much of this stuff, it's helpful to think about myself as with childlike playfulness, and this is I'm gonna have a lot of fun playing with this idea. And I'm I'm not gonna start running, but I do do a lot of walking. So I think on my walking, I'm gonna experiment with writing an article in a similar way that you described. So thank you so much for these both practical as well as broader ethical and philosophical things that you've shared today.

Leon Furze [00:41:26]:
Thank you very much, Bonni. Thanks for the invitation.

Bonni Stachowiak [00:41:31]:
Thanks once again to Leon Furze for joining me on today's episode. And Leon has shared a 20% discount code for Teaching in Higher Ed listeners to his practical AI strategies online course. So check the show notes for that information, and thanks to each one of you for listening. Today's episode was produced by me, Bonni Stachowiak. It was edited by the ever talented Andrew Kroeger. Podcast production support was provided by the amazing Sierra Smith. I'll see you next time on teaching in higher ed.