Higher Education
Explore what the future of curriculum design in further education should look like now and in the future, considering AI and Ofsted requirements
With the ever-changing environment of education, funding, demand, and technology, this session will cover what curriculum design should look like now and in the future. The discussion will center around using AI, critical Ofsted requirements such as employer partnerships in design, screening, assessment and learning technology to maximize innovation.
Presenter:
Ben Dainty, holds a Ba(hons) in primary education and a Master’s degree from Leeds Beckett University where he specialised in developmental psychology. He later transitioned to a career sparked by the sudden shift to online learning during the global pandemic.
Passionate about accessibility and quality in online education, Ben has previously worked at a national training provider where he founded and managed a team of teaching and learning specialists, publishers, graphics designers and videographers, progressing to Head of Teaching and Learning.
Now at FE Tech, Ben’s intense background in Psychology and eLearning development positions him as a powerful energy in the world of FE. Heading up FE Tech’s eLearning Academy as a Learning Scientist Consultant, he plays a central role in driving the content development strategy, supporting the FE sector whilst ensuring a seamless blend of innovation, accessibility, and quality.
Transcript:
Ben Dainty:
Hello everyone and welcome to the session that titled The Future of Curriculum Design. My name is Ben Dainty and I work for a company called FE Tech, who you’ll hear a little bit about during this session. So this session is all about the future of curriculum design, how I implement AI into my curriculum design process. It’s designed to have some really practical applications that hopefully, you can take away and start to use in your practice as soon as leaving this session. So what will we be covering today? So this session is split into seven areas. The first one is just a little bit about me, just telling you guys what I do, how I do it, then a little bit about where I currently work for and potentially how we might be able to help you out. Then we’re going to move on to why we’re here. So setting the tone and the purpose of this session.
Then we’re going to be moving on to the ADDI model, which is a design model for curriculum design and just explaining that and how I use it in my processes. Then we’re going to be going into curriculum design paths, so looking at some potential pain points of what we’re all currently doing in our curriculum design processes. And then we’re going to be moving forward into a theory which I’ve started to think about, started to coin term four, which is educators in the loop. So how to use AI efficiently within education. And then finally, we’re going to be going into ADDI in action. So I’ll be explaining the model of ADDI, and at each stage of the ADDI model, I’ll be explaining how I use AI to design my curriculums, speed up my work processes and achieve a nice work-life balance.
So, a little bit about me. So I hold a BA Honors in primary education studying Leeds, then went on to study a master’s degree in psychology. During this time, the pandemic hit and then all my tuition started to go online and it was terrible. People were just sending me reams and reams of information. There was no easy way to digest it. So then I started to think, how can I turn this into more digestible online education? So I started to develop some content on Articulate and I started to develop my own ways of building content. And I thought to myself, oh, I wish this was a job. I wish I could do this full time. And luckily for me, that is a job. It’s called instructional design, and it’s what I do now. So thank you pandemic, in a weird way. So during this, it really gave me a passion for online education with my degree and my master’s degree in psychology, specializing in neuroscience and developmental psychology. I became committed to accessibility and quality and e-learning, making sure everybody can access it no matter what.
I then went on to found and manage a team of learning specialists and national training provider here in the UK where I progressed onto head of teaching and learning. And during my time, I’ve designed and made over 100+ courses all the way from small CPD programs to level ones to level twos, level threes, level fours, and then some higher education programs in there as well. Currently, I work at FE Tech and I’m the learning sciences consultant here, and I also head up our e-Learning Academy. So I am responsible for all our e-learning, content production and our instructional designers. So a little bit more about me. I also am a division one basketball player here in the UK. Not the same level is in the US, but still. I love roller coasters and this is my little buddy Arthur, my little hamster that keeps me company whilst I’m building all my courses. He won’t make an appearance today though, so don’t worry.
So a little bit about FE Tech, who we are. So similar to my career, we was born during COVID. We saw a lot of people buying technology here in the UK and potentially all around the world. China, Sri Lanka, the US, buying technology and not knowing how to implement it in their practice. They were just panicking, we need a solution, let’s get something in. And the prime example of that is VR headsets, especially here in the UK. During COVID, lots of people bought VR headsets. They didn’t buy a platform to host the content on, they didn’t buy any content, just had headsets and you saw lots of classrooms just people putting them on and having a look at a Google VR image, and that was pretty much it. So we were founded as an organization designed to help further education providers find the right learning technology solutions that designed for our industry.
So we offer this for a free membership for advice and guidance. So we help all of our partners on a free bit basis if they’ve seen something and they want to know how to implement it. But we also have lots of different partners, lots of brand new learning tech that we bring into the market. An example of that would be GoReact. Obviously, GoReact is very, very popular in the US. You guys are very, very far ahead of us in adopting that. We’re helping bring GoReact into the UK market. I personally think it’s brilliant and there’s lots of vocational courses in the UK and apprenticeships being a prime example, and GoReact is going to be an excellent solution in terms of quality, accessibility, sustainability, and we’re helping bring that to the UK market. So that’s what we do.
So why are we here? So we’re all here because we want to have an impact in education. We want to have the greatest impact on the largest number of individuals. That’s why I’m here. So I used to be a primary school teacher and it’d be great, there’ll be 30 kids every year that would total through the door and I could make an impact on those 30 kids. However, I wanted to do more and that’s why I got into instructional design. That 30 kid cohort that I was teaching could now be thousands, tens of thousands, hundreds of thousands that have done my courses over their lifetime. And then secondly, the thing called GenAI, generative AI. It’s here, everyone’s using it. Are they using it correctly? Maybe, maybe not. But we want to learn a little bit more about it, especially in the UK there’s a real reluctancy to use it as prominently as I’ve seen in practice around the world, and we really need to get better at using it here.
So four quick facts on generative AI. So just in case you don’t know what it is, you’ve been living under a rock, let’s all learn together and then we can move on now when we’ve also got a base understanding of what it is. Well, sorry, ChatGPT was the quickest adopted technology ever. So Netflix took three and a half years to reach 1 million users, whereas ChatGPT took only five days to reach 1 million users. So it shows there’s an appetite for this. It shows that people want to use it, they want to speed up their processes, they want to play around with generative AI solutions, which is great market value. So the market value for AI tools is expected to reach almost 40 billion by 2025. It’s going to be a huge in terms of people creating it to make money. But not only that, it also means there’s a lot out there.
So there’s a huge market for AI resources. A recent estimate shows that there’s over 1.5 billion AI applications in use today. That’s not 1.5 billion just in our sector, that’s as a whole. But when you think about that 1.5 billion, there’s going to be a lot in our sector. And believe me, I’ve looked for a lot of these tools whilst engaging in this project that I’m at the moment. Also, behind me, which I’m covering up, is the new workforce. So AI potentially WILL replace around 800 million jobs worldwide by the end of 2030, which is a scary thought. Personally, I don’t think this will be in education anytime soon. However, if any of you are moonlighting as radiologists, you might be in for a slight shock as AI can analyze X-rays and diagnose medical issues just as accurately or more accurately than doctors. So yeah, food for thought.
So moving on, the education industry is behind, in both awareness and skill of AI. A report by Forbes in 2023 stated that 60% of staff have used AI at least once. When compared to other industries, this is a very low threshold of people that have used it, especially considering quite a lot of these people that used it have only used it just once. But it’s not only that, there is a generational divide in AI usage in education amongst staff. Staff under 26 show the highest usage of AI within their practice. And whilst that’s great, we’ve got some young guns coming through, showing and trailblazing the way for using AI in their practice, this can cause an issue once we’ve got… So picture this, we have someone at the end of their career in education, they’ve worked their way up, they’ve loved teaching, stayed in the classroom.
We’ve got all this bank of wealth of knowledge of all their experience, and they potentially move, they need to apply for another job. Then we’ve got this young gun, 25, new to the sector building content quicker than anybody else ever has. It’s going to be a tough one between who we’re going to hire. My props probably goes to the younger person, but this generational divide can cause issues with getting rid eventually of the really good knowledge that we need to retain in the sector. So it shows that AI is important to what skills and usage of AI is important to develop in our older and more experienced staff members.
And it’s not just for the staff at all either. So it’s also for students. Students utilizing AI will create a new achievement gap, and it’s not through lack of knowledge on the subject they’re studying. It’s more due to a lack of modern day AI tools that they’re implementing in their studying. So a study in 2020 showed that participants in a study that used AI in their revision, learning, research saw an improvement of 8.12 percentage points on average whilst utilizing AI as a study aid. So that’s not using AI to do the work for them. That’s using it as a study aid, which is the primary use for AI for students. So that’s 8.12 percentage points. Creating a new gap. This is the new digital divide at students having access to AI within their studies. And this study called then after that, for us to foster AI literacy in students and instructors, and without a technical background, 8.1% is huge. So it really shows that we need to improve our tuition on AI to our learners.
So, what can generative AI do for instructional designers like me? Well, it can create horrifying images such as this one, I’ve cartoonified myself. I personally don’t think it looks like me, but you guys can decide that. But more practical applications of this in instructional design, sometimes we have some crazy ideas of an image that we need to develop, create to illustrate an abstract concept before we might have to design that ourselves on Photoshop or Illustrator. It’s a labor-intensive process or we’ve had to scour Pixabay and Freepiks for an image that kind of represents what we’re looking for. Well now, with applications such as Dali, we can get our wildest dreams to us within 10 to 15 seconds.
So it also can create video. This is OpenAI’s Sora, so I’ll play that while I’m talking. I’m sure you’ve all seen this. This is one of the most exciting things that I’ve seen this year. Not sure what that says about me, but this is fully AI generated from a two line prompt. It’s pretty incredible. But the wider applications of this are immense for educators and instructional designers. I used to run a team of videographers. In that team, we had up to 18 videographers at one time. It wasn’t just expensive to run, it was time-consuming. They were all great people in terms of the speed that we needed to create content. Now, with such applications as Sora, when it releases later in this year, we’ll be able to create videos for prompts in our courses. We can use it as great prompts for creative writing. If we’re struggling to find something that illustrates our point, we can just do this with a two line prompt is genuinely so exciting for us.
And then also it can clone voices. This is something that I use all the time now. So I’ve trained a model on my voice that every time I do a session, every time I record a video, I chuck it all into this model and I now have an AI clone of my voice, which is pretty incredible. And I use that now for voiceovers, for scripts. It can also translate it into different languages. Recently, I made a course fully in Romanian and I put my voice in there translating myself to Romanian and I didn’t have to do any of the voiceover work. So if you are new to instructional design, this cuts down cost of getting good microphones. It cuts down time and effort. So I’ll just play this now and hopefully, you guys can hear it. And yeah, let me know if it sounds like me.
Speaker 1:
This is an AI generated clone of Ben’s voice. I hope my human counterpart isn’t boring you too much.
Ben Dainty:
It appears it’s consentient as well. So yeah, that’s an AI clone of my voice. So this is another video from OpenAI Sora. So again, I’ll display this to you. This is again from a two line prompt and just I’m asking it to generate a video of a birthday party. It’s pretty incredible stuff. However, it’s not without its faults. If we go and look at this again, if I ask you to look at what they’re doing with their hands, look at the way that the candles are all burning. And especially now once she goes to blow the candles, none of them actually react to that. So it shows that AI Sora, especially being brand new, fully isn’t there yet. So it couldn’t probably create things, I’m thinking like chemistry or engineering, things that are really accurate and need to be accurate for our learners, it won’t be able to do.
But if we look back, pardon me, if we look back just a few years ago, I think it’s like almost two years. This is what AI video creation looked like at the time. I’m sure you’re all familiar with this Will Smith eating spaghetti video. So if we compare that two years ago, two, three years ago to where we are now, we are making great strides forward. And yeah, there’s really practical applications for that within instructional design and within any new learning.
So, now we’ve got all these great tools, it’s brilliant, but it doesn’t mean that AI is being used correctly all the time. So a prime example of this, if you’ve ever wondered what Boris Johnson eating cake on the back of a camel in the desert whilst wearing the American football helmet, looks like, it’s that. And then if you’ve ever wondered what Garfield the musical would look like in real life, it’s this horrifying image. But there’s more practical applications for this as well around AI being used incorrectly. So in terms of the future of curriculum design, and often when people go to AI solutions, they go straight to ChatGPT as the main thing. In terms of the future, making things for the future and making content that is up-to-date, cutting edge and brand new, as you may or may not know, the last update on ChatGPT [inaudible 00:16:44] April 2023.
So we’re missing out almost a year of brand new theories, concepts, ideas that our learners are going to be missing out on if we’re using ChatGPT as our current and only source of information retrieval. And I’d also like to talk around this concept of using ChatGPT and using AI chat boxes like almost a Google service. So tell you a little story about my dad. So when I first got ChatGPT, so my dad is a chemical process operations engineer. So manual handling, does a lot of really dangerous stuff. And using ChatGPT is beyond him. But anyway, I sent him my log in and said, “Dad, you’ve got to check this out. Have a little play around with it and see what you think.” So he went and had a play around with it. I went back on the account and there was only two search histories on there.
One was facts about George Michael, and then the other was tomorrow’s lottery numbers. So it shows that people that are inexperienced in this field are using it as an information retrieval source as its main purpose. And in terms of new people coming to this and using it for that, it can potentially be out of date, especially with things such as legislation changes. Particularly here in the UK, we have lots of different legislation changes in terms of healthcare. And a prime example of this is, if you ask ChatGPT now what the current minimum wage is, which is updated in the UK in April, it will tell you last year’s current minimum wage. So if you are making a course on minimum wage and labor rights in the UK, if you pull your content from ChatGPT, it’s out of date. Prime example.
So moving backwards rather looking at the curriculum design past. So here’s a few images that I just want to talk about just to illustrate things that we’re doing now that can be solved through the use of AI. So this one on the left is just reams and reams of paperwork, curriculum design plans, storyboards, mapping, quizzes, homework. I’ve had desks this high, same with this. With AI tools such as Almanak and Formative, we’re able to store all of this and also not have to do any of the hard work in terms of writing everything down. This can all be sped up for us. The next image on the top in the middle, this refer to accessibility. So this is performed very passionate about accessibility and through AI now, we’re able to make content as accessible as it can be using the platforms we are.
There is AI accessibility checkers that can check everything for you and make sure that everything you’re putting out is accessible to learners. But then in terms of using text-to-speech software, speech-to-text software, all of these now are coming through to enable learners to access the information in the way that works for them. A prime example of this would be in auto AI, if you’ve not used it, have a go on your next Teams meeting and see if you digest the information more when you’ve got the transcript coming up down the side. I know I do.
The one on the top right represents four of them. And obviously, I’m not saying this is for any of you guys learners, you’re here, you’re trying to make a difference. So I very much doubt that’s what your lessons look like. But in terms of creating personalized pathways that are more engaging for learners through AI and through an LMS that can really model and structure personalized learning experiences based on learners interactions with your core content. So in terms of monitoring click rates, retention rates, we can start to personalize the programs that learners are doing in instructional design, which is brilliant.
Pardon me, sorry. So the one in the bottom middle, the big team meeting, that represents course validation. So especially in the UK with any programs that creates particularly a new form of course that we have here in the UK, which is called a skills bootcamp. They need to be validated by employers. So the whole point of a skills bootcamp is to take loans in. They have very intensive 12 to 16 week programs and the idea is to get them a guaranteed job interview at the end of it. So for this skills bootcamp, it’s got to be aligned to local labor market intelligence. It’s got to hit everything that that person would need to do a job within 12 to 16 weeks. And the way that we do that is through course validation panels such as the one on the bottom. I’ve done loads of these in my time over instructional design, meeting with key stakeholders in this instance, future employers.
And they can be very lengthy processes. You can have one or two key voices that are spitting out all the ideas to you and you’re not hearing from everybody. And now, and I’ll demonstrate it later how I do all my course validation, I can automate the full thing, I can get really good feedback from it, I can separate that into themes. I can have AI separate that into themes, and then also I can feed all that into something like ChatGPT with my course specification to make those relevant changes to the course without having to really lift a finger.
And then that final one is just around adaptability. Through AI now, we can present information to our learners in any form that they wish to be presented to. So for example, if we have some text-based learning that we’ve got going on, we can then also instantly provide them with an audio recording of that. Say you’re issuing homework to your learners, which is text-based. Now, you can also put a little QR code on there using a clone of your voice and explain the key points to them. Also, in terms of adaptability, you can give them a virtual tutor on a QR code now on their homework that will help them through that processes. And that’s something that I’ve started to do in my curriculum design processes including these virtual chatbot tutors. And later on, I’ll show you how I build them as well.
So, I set out on a journey really to make AI toolkits. So through using all the information that we’ve just seen, and especially the study by Long and Margiela in 2020 around the 8.1% difference between learners that use AI and didn’t use AI and the generational divide and staff, I wanted to create some AI toolkits. So some AI resources, e-learning content for staff and for students that distill to them how to use AI appropriately, what apps to be using, how they can embed it in their studies, how much time they’re going to be saving. And the idea for this is I really wanted to make these AI toolkits purely using AI, using the AI solutions that are suggested in the tool. Just making a full thing about AI, with AI, for AI is a very matrixy process. But anyway.
So I started off with this. So I inputted what I wanted back from ChatGPT. So I was asking write me some content on AI in education and how it can be used. And I really learned the hard way. So I got some sample content back from them and it didn’t feel right. It didn’t know my learners, no matter how much [inaudible 00:24:40] prompt I put in it, it didn’t know the sort of tone that I wanted to take, my tone of voice that I wanted to approach. So quite colloquial and friendly in the student version and then more assistanty and helpful in the staff version. It wasn’t getting that. I was really upset about this and I was like, oh, it’s not possible. I set out on this journey I wanted to first course made by AI, fully by AI, about AI and it wasn’t possible.
So I went off to basketball training one night with my friend Ross, he’s a senior developer at Sky Bet. And he said to me, he was like, “Oh, well have you heard of this theory of humans in the loop in terms of AI?” I said, “No.” And he went, “Well go and have a read and tell me what you think.” So I went off and read about this theory of humans in the loop. This is Ankit and Ankit is very high up on the development area of Barclays. And he said that the future is human in the loop AI driven technology. He says human in the loop is a mechanism where powerful systems are set up. So they allow both humans and technology to interact continuously. So in essence, humans being placed in the center of a continuous AI process, having a site of everything that’s going on, continuous process. And that really got me thinking about how we can apply that in education. And that got me to think about educators in the loop, which is written by me.
So Educators in the Loop. Educators, you guys are all the best at what we do and that will never be replaced, unlike radiologists, in my personal opinion. It’ll be enhanced through AI and [inaudible 00:26:26] technology, but we’ll never be removed. So we need to take the teachings as software developers and we need to place ourself in the center of that educators in the loop, that human in the loop to verify content for learners. So I started to build out this process of what it actually like. So input, we use generative AI to create our desired content. We need to verify that our input is exactly what we need it for. So AI doesn’t know our learners, it doesn’t know what we’re expecting back from it. So that input needs to be as specific and to the point as it possibly can be.
Then again, we need to verify again the content, the output that comes back to us. So we need to verify that that meets our expectations and adapt it as needed. For example, if we’re getting content back, and it’s reams and reams of texts we’re teaching in say a year six class in the UK, in primary school, that’s not going to be in a usable format. It can be a great start for us. But in terms of breaking that content down, understanding our learners, understanding how they like to learn, how they like to perceive information, that’s where again, we need to take a little bit of it and then verify that it meets our learners through adapting it as we need to.
And again, that comes in for our final stage, which is implementation. So we need to place the content into our curriculum in a way that our learners expect to see it. So if we’re again that reams and reams of information, does that need to be presented in a video? Does it need to be presented in an activity? Does it need to be presented in a role play scenario? We’ve got all these great ideas coming through from AI that are helping us and springboarding us onto our next step, but now we need to take control from AI and implement it into our curriculum in a way that is useful for both us, that matches our delivery and teaching curriculum style, but also matches the way in which our learners are expecting to receive that information.
So, in terms of educators in the loop, that’s what it looks like. Us in the center there, verify our input, verify our output, implement in the way that we deliver, and learners expect to receive the information. So good news is, I did it. I created these AI toolkits, one for staff, one for students, separate resources that not only showcase the best AI tools there out there, commonly found issues all the way from managing timetables, creating flashcards, creating curriculum, video production, all this sort of stuff. It also showed them how to do it, step-by-step guides and encouraged them to try to use them as well. So they’re one for staff and they’re one for students. They are available if you’d like to have a look at them. If you send me a message in the chat at the end of this, I’ll follow up with you and I’ll gladly grant you a demo to have a look and see what you guys think.
So ADDIE model. We’ll come on to the good stuff. One second. So A, A stands for analysis. So that who, what, where, when, how? What are we doing? Why are we doing it? Who are we doing it for? All those key questions and answers. The next one, D, stands for design. So this is an overview of the course of design included storyboards, any ideas that we’ve got from specifications, that type of thing. The next D stands for development. So this is where we’re developing our course content, we’re making our videos or we’re making images, worksheets, quizzes, actual course content, that comes under this section. The I stands for implementation. This is where your courses are live in your LMS or they’re live in your actual face-to-face tuition curriculum. And learners can start to take and complete those courses. And then E stands for evaluation. And it’s last but definitely not least it’s the most important part in my mind. This is evaluating what you’ve done and then create an actionable changes for the current course and future courses.
So this is the reason why I really will like the ADDIE model it’s because it’s continuous, it’s cyclical. That evaluation can then go straight back in and feed into your analysis process again. So if you are evaluating, you’re identifying something that isn’t working correctly for your learners or pass rates, attainment rates, interest rates, drop out, that can go back into your analysis. So maybe who isn’t right? Maybe what you’re teaching isn’t right. Maybe when your tuition hours aren’t right, maybe the way that you are delivering it, very vocational course if you’re teaching it very didactically, not very hands-on. Maybe that’s the why, the how you’re delivering it, which isn’t correct. So it all feeds into this nice cyclical process of you continuously updating and improving your curriculum.
So analysis. As we’ve gone through there now we’re getting into ADDIE an action, we’re seeing it, understanding what it is, and I’m going to give you some AI tools to go and take away and use to your heart’s content. So analysis, what is the purpose of the training? What is the purpose of the training? Why should we do it? What is the desired change that we’re trying to get through learners taking this course? So what stuff are we going to be doing? What difference are we going to be making at the end? What change in our learners are we trying to create? And then will that training be effective in creating this change? Is there any research that we can have a look at that has told us the best pedagogical devices to use in our curriculum and able to make as a change at the end of it?
So, in terms of analysis, one of the most important things is course validation. So course validation from students, course validation from employers. And the real value in that is building the curriculum that is purposeful and meets the needs of employers and the expectations of learners. Those are two key stakeholders here in the UK. Our learners and our employers. And if we satisfy the need of both of them, we can’t be going much wrong. So in terms of analysis, it used to look like this. So this is what stakeholder engagement or course validation used to look like. Sitting down with learners, what do you want to learn? How do you want to learn it? You often get only eight to 10 learners in one session. It would be quite slow, potentially learners could be quite timid and you wouldn’t get some really high quality solutions out of that.
So how do I use AI to do this now? Well, I use a platform called Typeform. Typeform is essentially an AI enabled questionnaire tool. It enables quick design, it’s got AI features which expedite that design, tailor it to your expectations. It converts already existing surveys. So say you’ve got a learner survey from your learner voice strategy that you send out to your learners at the start of a program or before they even join your institution, that can then just be automatically put in there and I’ll show you it in a second, how quickly we done. Under 10 seconds, it’ll create you a survey and you can send out the link to your learners.
And then the really good tangible bit that I really like in Typeform is that it identifies key themes for you in this tab called insights tab. So if learners are saying things that are semantically similar, but you’re struggling to read between the lines of 30, 40 responses, this will identify those semantically similar themes and break them down for you. So if they’re all talking about maybe like on-boarding or they’re worried about this or they really want to learn about AI but not saying it in that way, but semantically, they are, it will pick that out and present that information to you, which is pretty amazing. So let’s see it in action.
I’m going to share my screen and I will show you how I do it. So here is Typeform in all its glory and here are some surveys that I’ve created for some questionnaires. So just to demonstrate really quickly on, just to give you a guided idea of how I do it, there’s two that I like using which is Import Questions and the other which is Create with AI. So I’ll demonstrate Import Questions really quickly for you. So here, I have a survey, which is an AI course feedback survey. This is the survey that I use to figure out my loan and objective aims and goals for the AI toolkits themselves. So similarly, if you have a learner voice strategy and you’ve got the surveys that you send out to employers or you send out to your learners, they can be used using this. Type your questions in or copy them in, paste them in rather, and then press generate form.
It’ll then use some AI in there to figure out what questions you’re asking your learners or asking your viewers rather, and then identify the right question types for them. So here, it’s analyzed them and it’s placed them in the right form of questions. So multiple choice, open-ended, a sliding scale. There are extra things you can do by adding labels to the zero, five and tens, give learners a little bit more information. But you can also change the design for it as well very, very quickly. You can build your own themes down the side here. You can say, so some are paid, but let’s choose the brown one for example. You can change the look and feel of it overall. But then a really good bit from that, so if you don’t have a starting point, you can create with AI. So all you need to do is type what you’re wanting.
So a survey for learners on an AI e-learning course. Continue. What questions? Let’s go study modes, what topics, length of course, give it some ideas. Go continue and then for education. And then it’ll generate that form. So it’ll take a few seconds and automatically generate as a form in which we can share to our learners. So let’s see if this one’s any good.
So welcome to our study. Which study mode would prefer online, in person, hybrid, and we can add all these questions as well. What topics do we want to learn about? How long would you like the course to be? So we can add them, we can add endings. But then in terms of sharing, it’s really easy. You press this publish button up here and you can copy a link. That link can also be embedded in iframe. So you can embed that in emails in a website as well, like an enrollment form. And then also, if you put that into something like QR Code, Monkey, you can check QR code form to be sent out on flyers at open days of perspective events to get more and more information about how you should be altering your curriculum. So I will stop sharing that screen one second and head back to my main screen.
So now, we’ve got what our learners are expecting to see, what they’re expecting to learn. How can we then use AI to assist us in our curriculum design process? Well, a really easy example and something that I do quite a lot is using ChatGPT to develop our curriculum plans. So taking from our surveys in Typeform and the curriculum that we’re trying to deliver for our apprenticeship or our qualification, anything, we can put both them into ChatGPT to put in your survey results or in your curriculum that you’re trying to deliver or the specification for your course. And then simply prompt ChatGPT to say something like, I need to combine some actionable points from the survey into my curriculum. Can you suggest relevant points in which I can alter and adapt what I’m going to teach my learners? Works very well. Have to play around with it to figure out if it works for your curriculum, but I’ve used it to some great success.
Next or alternatively, you could also use Almanak. So Almanak is an automated content generation tool. I think it’s actually purposeful built for the US actually. So if you’re not using this then it’ll make more sense for the audience that are from the US to use this. It can use your own existing lesson plans or AI generated lesson plans to plug into one of their generators where you can create slides, content, resources, quizzes, report cards, and it can also be integrated into your current LMS or CNS as well. So it can speed up processing times. But the one that I really like to use is ID Assist. ID Assist is an instructional design assistant. It helps me when I’m stuck with how I’m going to shape my curriculum or what things I’m going to add. Yeah, I really like this tool.
So let’s go and show you what that looks like. So I’m going to need to stop sharing on my screen in a second. So here we have a piece of content that I’ve been writing and this is for a course in the UK. It’s an education training course at level three. So this is just some bits of content I’ve been writing all about the first aid assessment criteria at the top, which is understanding the teaching role and responsibilities in education and training. This is all about understanding items around that. So ID Assist is an extension, it works in Google Docs and it’s also just launched in the Microsoft packages as well as a word and PowerPoint. So once you open it opens up in the right-hand side here. It has lots of different instructional design prompts that you can use. And the reason why I like using this is it keeps you in one place. You’re not taken away from your content to alter a little bit part of it.
Like if you were plugging these into ChatGPT, you’d be flicking back on board. Even if you’ve got two monitors, your focus isn’t on where it should be, which is how the content looks and flows on [inaudible 00:41:04] document. So here’s just a few prompts that we’ve got down the side, animation scripts, we build FAQ scenarios, that type of thing. So if we go into this first one here, this is all about what the role… Actually, let’s go for this one. This is facilitating learning. This is one of the key responsibilities of teachers. So let’s build an analogy for this one. So we highlight our text, select our prompt, generate content, it’ll then read that and then develop an analogy for us to explain it to our learners in a form that they might understand a little better. And here it is.
So imagine a teacher as a chef in a kitchen. Chefs may not always [inaudible 00:41:49] cook food, but to ensure the meal is delicious and nutritious and caters to the preferences and dietary needs of the diners. Similarly, a teacher’s needs to learners, it’s not just to deliver information but to facilitate learning by making the content engaging, accessible and suitable for the needs of diverse needs of students. Brilliant. Excellent. I’ll take that. Thank you very much. And I’ll pop that down below and then incorporate that as a part of my content. I can even use say like [inaudible 00:42:19] for example, and create a nice image to represent something like this. Let’s do one more. Let’s come down to the bottom here. Let’s go for professional development. So here, I’ve given some example, actually I’ve not given any examples. Let’s change that. Let’s go to a custom prompt and let’s replace this prompt. So provide me with a short list of examples of professional development.
If I could type, that would be brilliant, wouldn’t it? And then we’re going to generate. Excellent. Got 10 examples there. We’ll take that, place it below. And now, we’ve got some actionable examples for our learners to understand different CPD activities. Obviously, going back to educators in the loop, I’m going to have to look through this at a later date. I won’t do it now, but just verifying the accuracy of the information in there using my knowledge and experience. So I will stop sharing my screen and open back up the PowerPoint. So one second. Excellent. So now, we’ve done our design and now we’re moving into our development phase. So an example of an AI tool we can use in a development phase is Formative AI. So this is an automated solution for grading homework or tests. It can automatically build them for you, automatically mark them for you and then release your results to learners. And that’s all data track. So it gives you great insights to the progress of your learners as well.
But moving forward, here’s something that I love using. The reason why I love using this is when we was checking engagement of videos and usage in our courses through monitoring click rates, learners would look at a video and go, great, I’m off for a quick tea, this is a break for me, I don’t need to pay attention. And then would see some drop off for like 20 minutes, half an hour when a video came on, they’ll just go probably watch a TikTok or who knows. But anyway, Edpuzzle solves this issue and it really brings the power of video learning. So it transforms videos into interactive and gamified experiences, making sure that your learners are paying attention, they are engaging with the resources because often, especially in the videos that I create, we embed really important information in those videos, case studies and scenarios and interviews with real-life people who are working in the industry they want to get into.
So let’s go into Edpuzzle and just take a little look at it. Here’s what Edpuzzle looks like. It is a web-based solution. And you can create classes, so you can have different classes for your different classes that you’re teaching and keep everything all nice and tidy and new. So once you’ve found a video that you like, and it can be your own video as well, it doesn’t have to be anything from YouTube, but you can search that in a bar at the top.
If it finds it will then present you with a timeline like this. If it’s your own video, you can add your own voiceover to the video, which is great. So in terms of accessibility, you can add points to personalize your classroom. You can also add instructional points in terms of reading the questions out to your learners as well.
But then let’s get to those questions. So this is where Edpuzzle comes into its own. So you can add your own manually multiple choice questions and open-ended questions. You can also add notes to your learners to personalize it a little more. So if someone’s talking about the development phase of the ADDIE model like we are now, we can add a little note saying this is what we’re learning right now, bring it to life. But then what it really comes into is the teacher assist, which is still in beta, but it’s still working great for me. So here, it’s going to use AI to generate questions based on the transcripts of the video and place them into the video so we can have open-ended multiple choice or both. Let’s go for multiple choice for this one.
What it’s done now is analyze the transcript and place four questions along the timelines where these little purple droplets are. So, straight after the knowledge has been disseminated, we can then answer a question. So, we click on one of those little purple balls at the bottom. The video will pause and when it gets to that, learners will be presented with a question. So which of these is not a phase in the ADDIE model for instructional design? But yeah, it’s that top one and it’s been marked as correct there to show us in terms of it being correct. [inaudible 00:46:56] want to edit it, we can come in and then change the right ones if it’s not particularly got it right. In terms of sharing this, it can be embedded within any LMS. It can also be sent to your learners on tablets or on laptops through QR codes. You can send them emails to it. And also in terms of tracking, you can track it all in your Edpuzzle program as well.
Let me just stop sharing my screen. Cool, all right. So now we’re on implementation. So implementation is all about implementing your content and your curriculum design into your LMS and building a lot of appropriate things in there to help your learners progress. So for this one we’re going to talk about Zapier Interfaces. So this is all about building an intelligent tutoring system to assist your learners with their progression. So Zapier allows users to create a new thing called custom interfaces, which can be developed into dynamic and tutors for your students. It can be designed to supplement in-classroom learning by providing immediate feedback and also out-of-classroom learning if it’s designed for more personalized assistants. So we’re just going to go in and have a look at my example of this one, which is Educhat. So let’s bring him to life as I adapt my screen slightly.
So, here’s Educhat, and this can be embedded again in iframes, LMS, you can enable access to this via QR codes. And the way that this has been coded, it’s been coded to tell learners to scaffold learning, rather. It will never give learners the answers to the question unless they’ve provided the answer for that question. So it’s perfect to help pop a little QR code on top of homework, learners scan it, say if they’ve not got access to people at home that can assist them with it, they’ve got something on the phone that can help them do it.
So let’s just go, what is seven, eight, nine? Let’s see what it is. So, this expression, you need to follow the order of operations. I did switch this to the US, I think this is US PEMDAS is that parentheses instead of brackets. So it’s going to ask you to start identifying what operations we need to perform first in this operations. So we could go… Terrible spelling. Sorry everyone. Yeah, there we go. They come first in order up. So it’s scaffolding our learning and that’s the whole point of this, just the scaffold learning.
So finally, we’re now moving on to evaluation. Evaluation again being a really important part of this curriculum development cycle. And I use ChatGPT to analyze my data because it enables me to develop little insights into it, ask it for responses, how I can improve. I really like it, it works for me. So let’s see it in action. So I’ve got a little example here.
Okay. So essentially, I’ve inputted my spreadsheet of results. My spreadsheet of results is across lots of different assessments and it shows the learner identifier. So learner number, the score for the assessment and an engagement score for the assessment. So learners self-reported how engaged they were throughout the teaching, an assessment of that unit against the score. We can see if there’s any correlations in there. So once we’ve been inputted that in there, ChatGPT has then asked us if this is correct. It’s essentially said, learner number, the score assessment and the engagement assessment. If that was wrong, at that point we could then edit that and tell them that certain identifiers or labels it’s identified are wrong and get it in the right way. So we can ask you questions about the data. So instead of going to your Power BI team, if you’re not data people or looking through the data yourself, you can just use ChatGPT as an assistant for you.
So I wanted to know if the engagement score for each assessment correlated with the assessment score. So ChatGPT came back and said the correlation between engagement scores and assessment scores for each assessment is strong as you would expect. We’d expect that. And showed me our Pearson coefficients down the side, showing how strong it was. I then asked, because I’m a visual person or I’m a visual learner, can you create a chart to show that? And it showed me this bar chart showing the efficient Pearson correlation on the side and the assessment. So relatively high. Now, I want to get a little bit more granular and I want to know who had the lowest score overall, who do I need to identify earlier in this process? So it said the learner with the lowest score was learner four and had an average of approximately 63.43% across all assessments.
And then I was like, well where did it all go wrong for this learner? Where should I have identified it? And I said, is there a point throughout the data in which they start to drop off? And it said, “The problems of learner four across the assessments are shown in the charts and the chart above.” By examining it, we can identify it. So by assessment two, their [inaudible 00:52:14] a bit high and then they dropped in assessment three, and it’s a bit of a spiky profile, but what that can tell me is that they really didn’t like assessment three or the teaching for assessment three, or maybe there’s something going on in their personal life around assessment three.
So from that, I then asked ChatGPT to be a data expert and provide me with 10 useful insights into this data. So it provided me with 10 useful insights for me to start my cyclical process of evaluation, feeding back into analysis that A at the start of the ADDIE model. So obviously, it’s strong correlation between engagement and assessment scores, but there was varied performance across different assessments. This variation indicated different levels of difficulty of the assessments. I could review that.
Learner four’s performance drop shows a notable performance drop in assessment free. That could be a specific area of difficulty or external factors as I said before, highlighting that. And there was an issue with consistency in high engagement. So what could I do to re-evaluate the how in ADDIE or the why, can I change in my curriculum process next time? Then there’s a good thing in here. Assessment six was a high point. Assessment six shows the highest correlation between engagement and scores sharing. It’s particularly effective and aligning with my teaching methods. So what I could do is go into assessment six, look at the content and what I’m teaching, look at how I’m teaching it, and then maybe take some pedagogical skills or styles from that unit, take it back to the analysis section and start to implement those across my other units and see if there’s any changes that I can be making.
So that’s everything today. Thank you very much for turning up and attending my session. I really hope that you’ve enjoyed it and you’ve got some practical examples of using AI or you now understand a little bit more about something. The QR code on the screen is a free to scan, free to sign up. Obviously it’s free to scan, but sign up to our free membership here at FE Tech. If you do have any questions regarding tech that you are looking into and you’re not sure of the quality or you’re looking for solutions that you’re not quite aware of yet, that you’ve got a real niche tech gap and you need some help with solving that. Sign up to that and then we can always help you with that once we get in touch with you.
So thank you again for attending. Much appreciated it. If you put any questions in the comment section down the side, I’ll personally respond to each one of those personally via email. I’ll give you proper and full response once I’ve had time to think and digest what it is you’re saying and I’ll follow up with you later. So once again, thanks very much. Goodbye.