ETMOOC 2.0 Session #2

Welcome everyone to the second session of ETMUC 2.0. So happy that you could join us this evening for panel presentation. As we get started, I would like to welcome you and also share a land acknowledgement. And I just want to share the image that I provided here. I was playing around with AI Now I can't remember which tool. It seems like there's new tools coming out every day. And I used one to play around with this image which didn't have a body of water in it. And so I asked for some of the land to be replaced with water and for the mountains to be changed a little bit. I'm not sure I really improved the image, but thought I would share that with you as I take a moment to acknowledge where we work together to advance teaching and learning for all learners on the traditional territories of the People Treaty 7 and Metis Nation of Alberta Region 3. And that's where all of us as panelists are located today. I thought I would share that with you. Great. And if your audio is not on mute, you may want to put that on mute at this time so that we don't have sound interference in the background. So I'm going to take a moment to introduce our panelists this evening and also wanted to share with you that I created these slides using, I think it was Gamma was the. The tool. And I just put in what the title was for our presentation. And I'm not going to use all of these slides, but happy to share them with you. It went through and created definitions, benefits of AI challenges, AI tools in education, the future of AI in education. So some of the things we are going to be talking about, I just thought I share with you slides that I'm not actually going to use.

Instead, we're going to use some questions.

That we're going to ask. Okay, I'm still hearing some sound in the background, so I don't know, maybe one of the co hosts could help us with the meeting. Thank you, that's great. All right, so I'm going to begin by introducing our panelists. And I know they have their videos on, so I will just ask you to wave when I introduce you. So for those of you that don't know me, I'm Barb Brown and I'm from the Workland School of Education at the University of Calgary. I'm an associate professor in Learning Sciences and also the Associate Dean of Teaching and Learning in our faculty. And I'm so pleased to introduce to you the panelists that are here today. So we have Dr. Sarah Eaton. I'll just wave. Thank you. Sarah, who is associate professor also in the Wirklin School of Education. And she many of you, you're probably all of you in the room may know her as she's an internationally recognized researcher and scholar that focuses on academic ethics in higher education. Dr. Soroush Sabha Gat is also my colleague from the School of Education and Director of Equity, Diversity and Inclusion with research interests in artificial intelligence and addressing pedagogical challenges in higher education. So pleased that you could join us and my colleague, Dr. Brenda McDermott who is the manager with the Student Accessibility Services at the University of Calgary with a focus on accessible testing and so glad you could be her Brenda and join us as well. So as you can see, we have individuals who have a variety of expertise and interest in related to artificial intelligence. And our topics tonight, ethics, equity and accessibility. And so we're going to ask our panelists some questions around these topics. And as we do that and as you hear their responses, please feel free to jot any questions you might have in the chat so that you know, we can ask them or continue to engage in the conversation or you might want to think about. I know in the on the minimooc website there's also a section for our session where we encourage you to blog or continue the conversation in discord or other social media that you're using, using of course, the hashtag for the so that we can continue that conversation. So at this time, just right before I get the panelists to start answering questions, it might be really helpful for them to just get a sense of who you are. So just really quickly, if you could in the chat, just give a mention your role or where you're coming from and then that'll just give us a sense of, of who's in the room. I'll just give you a moment to do that. So Susan, psychologist. Thank you, Susan. Librarian, coordinator and coordinator of ict. Someone asked me the other day, what is ict? I couldn't believe it that, that that was a question but not, not necessarily an acronym that maybe everyone is familiar with. So grade eight mathematics. Yes, I used to teach grade eight mathematics. Lots of educators, lead tutor, mentor, writing instructor. Paul from San Fran. Thank you. Writer, presenter, consultant, instructional designer, director, teaching and learning center, educator, staff. Okay, so as you can see, we have a lot of expertise in the room and so let's put our heads together and talk about AI ethics, equity and accessibility. So Sarah, I'm going to start with a question for you. I'm wondering if you can start us off by sharing your perspectives related to ethics and artificial intelligence. I'm wondering if you could tell us how you define ethics relative to AI and perhaps what are some of the ethical considerations that folks in educational contexts should be thinking about.

Thanks, Barb. Thanks for the invitation to join you today. It's great to be here with everyone and always happy to talk about my favorite topic on academic integrity and ethics. You know, there's no universal definitions of academic integrity. There's different definitions that exist out there. Often they exist within institutional policies and they're often framed in terms of, you know, behaviors that we don't want students to do. So integrity is often defined by behaviors that are also defined as misconduct. Right. Or is the opposite of misconduct. But for those of us who work in the field, it's usually about more than, you know, just avoiding plagiarism or not cheating on exams, but rather academic integrity as a foundation for ethical decision making in and beyond the classroom. So connecting, for example, academic integrity to workplace ethics and everyday decision making can be a really useful and constructive place to start having conversations with students about what it is and why it's important. I think if we just frame it in terms of, you know, avoiding behavior or avoiding punishment or avoiding consequences, it doesn't really help with student, student learning. So those are kind of some, some of the big pieces around it with regards to its application. With regards to artificial intelligence, I think there's a lot of uncertainty. So a few months ago when ChatGPT was just released, people were messaging me on social media and sending me emails saying this, is this a plagiarism tool? We're going to have a rampage of plagiarism. Right? And I think it really didn't help when folks like Noam Chomsky weighed in and said that artificial intelligence writing apps were nothing more than high tech plagiarism. That for me was kind of heartbreaking to hear that from such a high profile linguist because I think there's also lots of opportunities. One of the resources I wanted to kind of share and highlight with folks today is when we shared on the website, it's brand new guidance like just out this week from the European network on academic integrity. And it's around the guidance for the use of AI in education and how it might be defined in policy. So they're looking at possibly the introduction of a new term like unauthorized content generation. I'm not really sure how that will get unpacked. But also proposing things like training for academic staff, educators, ongoing pd. So I think to answer your question, we're still figuring out what some of the definitions are.

Thank you, Sarah. And just to continue along the lines of kind of defining these terms, Soroush, I'm going to turn to you next and wondering if you can share your perspective related to equity and artificial intelligence. So how would you define equity relative to AI? What are some of the challenges that educators may face when integrating AI into their teaching practice? And how can they be overcome?

Thank you, Barb. Good evening everyone, and thank you so much for having me. I think it's very important to recognize that equity is a multifaceted construct with different meanings depending on the philosophical perspective from which it is derived. So I'll provide two examples to clarify. So consider utilitarianism, which is about maximizing happiness. How would equity in the healthcare system be defined from this perspective? It would mean allocating resources to those who are likely to benefit the most, such as individuals with the greatest needs or those who are most likely to survive. Now consider a deontological perspective which is grounded in moral principles and duties. And in the same example, equity means providing additional resources to individuals who face systematic barriers or systemic barriers, such as people from marginalized communities or individuals with disabilities, regardless of the strain that it would put on society, because it's a moral duty to do so. So both of these perspectives believe in equity, but they have very different processes in achieving equal them. So going back to your question, Barb, I'm going to relate equity to artificial intelligence through a multicultural, intersectional and social justice lens because they're just more prevalent in Canadian context. Multiculturalism values the coexistence of mutual respect of different cultural identities within a single society. So equity means ensuring that individuals from different cultural backgrounds do not face unique challenges and barriers that limit their access to resources and opportunities. So from this perspective, AI should be designed to value and respect diverse cultural backgrounds. So, for example, if you ask ChatGPT to generate reading material for your for your class, let's just suppose that's one case of ChatGPT. One example of multicultural AI equity is ensuring or instructing the AI to generate culturally diverse reading material that prevent various cultural background that present various cultural backgrounds and perspectives and experiences. Intersectionality highlights the importance of considering multiple dimensions of diversity such as race, gender and class. So when promoting equity and inclusion. And it also talks about the need to address systemic discrimination and power imbalances. So an intersectional approach to AI equity involves different social groups. Again, suppose that you ask CHATGPT to generate reading material for your class Further, suppose that you have the AI produce material from various cultural backgrounds. One example of intersectional AI equity is ensuring or instructing the AI to produce diverse materials that do not systematically discriminate against marginalized or underrepresented groups and taking into account multiple dimensions of diversity. For example, last one so social justice theories focus on addressing systemic inequalities, empowering balances to create a more just and equitable society. So from this perspective, promoting equity means challenging or dismantling oppressive structures, ensuring opportunities, and promoting the well being of marginalized groups. So once again, suppose that you ask ChatGPT to generate reading material for your class. Suppose you've ensured equity from a multicultural and intersectional lens. An example of social justice AI equity is ensuring or instructing the AI to generate reading material that addresses issues that help students develop the critical thinking skills to challenge and dismantle those power imbalances. So there's no really easy solution to ensure equity in the design and development and deployment of AI. We know that interactions with language models today will be used to train language models of the next generation. So incorporating diverse perspectives and voices into interaction with these models such as ChatGPT3 can potentially contribute to creating a more equitable language to more equitable language models in the future. And we shouldn't stop there. Large language models have feedback mechanisms and when we see issues related to bias and discrimination, we need to report them. And combining the two methods, interaction and feedback, I think it's possible to create more equitable large language models in the future. And I hope that I was able to answer some bits of your question for sure.

Yes, I like the optimism. Thank you so much. All right, Brenda, I'm going to ask you a question just to give you an opportunity to kind of situate your area and wondering if you could share your perspective related to accessibility and artificial intelligence. So how would you define. And I know we've seen some messages in the channel already, it is really difficult to define these terms and maybe not even necessary to have one definition. But I think it's interesting to hear how you define accessibility relative to AI and what are some of the different interpretations perhaps that you've come across, Brenda, related to this term? What are some of the potential drawbacks or limitations of AI in assessment and accessibility and how might they be addressed? So it'd be an opportunity to introduce yourself to the group.

Wonderful. Thanks so much for having me. And it's great to follow Soroush, who's covered so much ground and allows me to sort of define accessibility very Much I think more from a user perspective rather than from the system's perspective. So often when I think about access, we've both kind of that who gets in the door, right? Who has the ability to use these tools? Whether it be, you know, we're looking at things along access to technology all the way to, you know, your ability to use it and its interaction with other tools. I also think about the benefits from using it and you know, do certain users have the privilege of being able to read the screen context easily or type their responses into ChatGPT versus those who may be needing to use another layer of adaptive technology? And maybe those interactions don't happen the same. So when I think about accessibility, I think about it in terms of does everyone have the same opportunity to engage with the technology and do they receive the same benefits from that experience? Historically, often individuals with disabilities or who are interacting with technology through layers of screen readers or dictation software or other adaptive tools have not received the same benefits. But when we hear the praise of potential artificial intelligence, it's often these groups who are being the ones who will be served best by this. So it's that kind of inclusion metaphor. One thing that I always find a little bit challenging when we think about AI and accessibility is again that variation and how the norm or those dominant groups tend to end up getting the benefits in the end. And I always find that a little bit challenging. And in terms of assessment pieces, AI and accessibility, it's that fear driven response that I always like to highlight. How often one of our reactions to we're not sure about a new tool, we think it might be perhaps eroding some fundamental skills. We create very restrictive learning environments. I think of the way in which we sometimes moved to online proctoring software and the groups who are often challenged, who face new barriers from these restrictive environments are those minority learners or those learners, more diverse learners. So we're adding, sometimes our choices not only try to protect us from our fear of the new technology, but actually reinforce and perpetuate new barriers for a large number of students. And that's often where I come in is we may need to qualify that fear or uncomfortable let we all of a sudden now perpetuate and expand barriers for learners.

Thank you, Brenda. So that gives us a bit of a start and gives you an introduction to the panelists and you know, their perspectives, where they're coming from. I appreciate seeing the questions in the chat and I know, Sarosh, you've already responded to one. Just wondering if there's, you know, Another question maybe, you know, Alec, I'm going to pick on you. I see your question in the chat. I'm wondering if you'd like to ask it or, or you'd like me to ask on your behalf.

I can try. I, I, you know, the wording is difficult sometimes. We know that ChatGPT, the, the dominant narratives, you know, the, the training data is from a dominant narrative perspective because all of our materials essentially are of a, of the dominant narrative and certainly there will be, you know, content developed from, you know, those who are marginalized. But of course, the, the bias, even if ChatGPT is able to hide or, you know, put, put under the covers that some of the bias, it's, it's lurking underneath. And so the decisions, what it, what it chooses to show us is always there. And there's a good example a while back that, you know, if you ask it in, in the early days when it was just released, I think November, December, if you ask it like, what are the characteristics of a good scientist, it would say, you know, it would give you a good answer. But if you did it through code, it would say if, if X equals M male and Y equals, you know, white, then that's a good scientist. So, like, from a mathematical perspective, it couldn't hide its bias or, or just, it's biased or discrimination from that. So it's really interesting that the language just cloaks some of the bias that's there and it's just a bit of an issue. So my question essentially was, can we ever see, even if we perceive it to have developed content for marginalized peoples, can it really, can we ever really trust it as, you know, as much as we, we understand that the content can't be far from being developed in a way that is not, you know, necessarily written from a, from a perspective of someone who's truly marginalized. I don't know if that's a question or not, but I should have just read the one in the chat. It's probably a little bit better.

No, no, it was, it was good. I'm wondering if one of our panelists might like to respond to that in terms of, you know, what are your thoughts and in terms of bias, I mean, I can, I can give an example the, the panel this evening I started putting together the plan and, you know, what questions I would ask the panelists. And, and then, of course, as I'm putting it together, I'm thinking, ask ChatGPT and see, you know, what kind of questions would, would I ask or could I ask panelists with this expertise. And so, so then I was provided with a panel plan and some questions that I could ask. And so then I looked at that against the questions I had written and, you know, and combined and merged some of them together and, and added some new ones. So it gave me some new ideas. And I put that into my blog, you know, in terms of the prompt I used and the output I had. So it was, you know, it was interesting to merge the two together, you know, in terms of bias. I think I had bias in terms of the questions that I asked based on, you know, my experiences and what I thought I should ask. And I would argue that the output that I received also had some bias within it. So it was interesting to merge the two. I'm not sure I was able to get rid of bias, but I acknowledge that it was there. So I'm going to turn it over to the.

Add something, if you don't mind.

Barbara?

Yeah, you know, even from a decolonizing perspective, if you think of, well, like, you can't opt out of this. Like, if you, if you do opt out of a tool that's this powerful, you're deceiving your power to, to a, to a dominant culture, to a techno culture. Like, there's, there's a, there's a great layer of, of technical techno colonization within this tool like that. Like, you can't avoid it, you cannot say no to it at some point. And then if you do opt in, you're subject to a tool that, that discriminates against you. So, so it's like there's no winning here in some cases. And, and I guess the other underlying thing is we, we really can't control it to the extent that we think we have. I mean, the big. One of the biggest powers problems of, of AI is the control problem. And so without being able to control what it says, what it does, how it acts, we'll never necessarily solve that problem. And so again, so I'm not meaning to ask questions that have no answers, but I think these are concerns that we have to think about. And I'm not, certainly by any means, someone who's a naysayer that we shouldn't use these tools. And I'm quite on the opposite side. I'm probably, you know, someone who is quite excited about the possibilities, but at the same time, we can't ignore these pieces. And, and certainly we need to work in unison with our students to not solve these problems, but make them aware and to critique the output, to critique the, the, the impact of these tools on the academy, but of course, just in. On society in general. Again, sorry, sorry for not having a question because I just don't know like these, these are questions, but I just don't know how to even ask these questions because it's so, it's so new and fresh and I. Those, of course, these things have been around in other forms for such a long time, but this just becomes so incredibly apparent and explicit that it just, it's hard to, to avoid questions like this, I think. You know, Alec, let me pop in.

With what I think would be a.

Good variation on the question you're asking here, which is what do we do.

To assure that our students are comfortable.

And become efficient, if not proficient at using AI? That sort of gets to the heart.

Of, I think what many of us are talking about tonight. Thank you, Paul. And I think, you know, it also, it gets at what we were hoping to, to accomplish during this panel as well is to, to have folks consider, you know, the uses, the possibilities, the, you know, the opportunities, but also the barriers and threats of, of using AI. And so when, you know, when I was thinking about the learning objectives for the website, you know, around the session, I thought, well, really like we're just, we're asking you to consider, you know, to reflect on and to think about. And I'm not sure we have, you know, single one way or one answer to any of the questions asked. So I'm thinking that maybe some examples might come up as we, as we start our kind of next round of questions. So, Sarah, I'm going to ask you the next question. So is it misconduct if using AI? You know, what are some examples of AI related misconduct in educational settings and how can they be prevented? Maybe through this you might provide us some examples, tell us about your research and implications for teaching. How do we help instructors, learners, school counselors, any role in educational context, move beyond the narrative that AI, artificial intelligence, that is equals cheating?

Yeah, I think the first and most important message is that the use of artificial intelligence tools does not automatically constitute misconduct. Right. I just need to kind of put that out there right away. You know, technology, any technology, can be used for good, it can be used for evil. I think a lot of it is in how it's used. And I think actually, Barb, you gave a terrific example when you talked about how you interacted with the AI writing tools, the AI tools to help you prep for tonight's session. That for me is a perfect example of life in what I've called a post plagiarism world where human and AI text becomes hybrid to the extent that we may not be able to tell which part of it is generated by a human, which part is generated by an AI, and really, why should we care? If the purpose of education is to facilitate student learning, then that's the. That could be our focus, right? There's a lot of time wasted by educators everywhere in trying to hunt down cheating. And I have a colleague in the US who says, you know, how can we reframe the driving question from why are students cheating? To why aren't students learning? So I think that's a really key question is if you know, how do we know students are learning? How do we provide opportunities for students to demonstrate their learning? So you asked for a concrete example of how AI apps could be used as a form of misconduct. Well, if students are are using, say, ChatGPT or any other AI writing tool to generate a response to a question, and then they cut and paste that question verbatim without looking at it, without reflecting on it, without interacting with it in any way, simply to complete an assignment which is effectively transactional, then that could be considered a form of misconduct because the students aren't learning. So if we think of this as subtracting learning from the equation, then we can say, well, we might think about, could cheating be inserted. I've also heard of instances of students using ChatGPT to generate responses to take home quizzes. So a big question in my world now is, well, if an assignment can be done by ChatGPT, is it really a good assignment? Or if an exam can be completed by ChatGPT, is it really a good quiz or an exam? And so these are big questions for us as educators. I think it has a lot less to do with students cheating and a lot more to do with us developing assessments that allow students to shine and showcase their learning first and foremost.

Thank you, Sarah. I'm going to have Suroosh, maybe build on that a little bit. Tell us about how you're helping faculty soroush plan learning activities in their courses and specifically around your portfolio. Consider equity, diversity, and inclusion. In what ways can AI be leveraged to promote equity, diversity and inclusion in education? And at this time, I invite you to screen share, if you like, and share the application that you've created.

Thanks, Barb. I think there are many different ways that AI can be leveraged to promote edi. One way that everyone's talking about is, of course, content generation. And I know there are. There are many instances where I've used AI to generate diverse, inclusive and culturally relevant educational content. And you have to do it by prompting the right way or chain prompting the right way. But one of the ways that is not often discussed is that large language models are also really, really good at identifying bias. Again, if you prompt them the right way and if you do prime training, they can pick up some very, very interesting things. But I'm going to leave that for a different time. I'm going to talk about using AI to support instructors and teachers, specifically in the domain of equity, diversity and inclusion. And I'm going to show how AI can be used to help teachers design EDIA Aware learning activities and assessments. Which brings me to this application I've developed, which is called Smarty. I'm going to put the link to it into the chat. I'm going to share my screen for you guys to take a look at it. Let's see, we'll do this one. So Smarty is basically a student of applications, is powered by ChatGPT and Anthropic's client API and it's designed to help educators create inclusive course outcomes and outlines. I'm sorry. And it has different components, but the ones I'm going to focus on is with regard to edia. So let me give you an example. This is a course I taught last semester, graduate level course. And in week 12 we did an activity which involved use ChatGPT to design the lesson plan. That was the objective of the learning activity. And I gave it students very clear, I think, instructions on exactly how to go about doing this. So I'm going to copy these instructor instructions. When I go back to Smartie, I'm going to go to an application that redesigns learning activities and then paste in the instructions. I'm going to say my course is asynchronous, I'm teaching graduate students and I want to redesign for multicultural students. And you can choose all of these, but I'm just going to use one because we're short on time. So what's happening now is chatgpt is reading the instruction instructions and looking for ways to improve this. So that's the first call. The second call is after it has identified ways of improving it, it's going to identify where to improve it. Like where exactly in the instructions should it make the alterations. And then the third call is what type of changes must be done to the task for it to be improved. So it makes three different API calls and it comes to this table. It has recommendations where to implement and how to implement. So let's see here it says use culturally responsive teaching and strategies where throughout the inside lesson plan, especially in the introduction and guided practice sections. And it talks about research and incorporate diverse perspectives and experiences. I'm not going to go through the entire. You have the link, I have some funds. I've released the app and I'm hoping that you would play around with it and see what it produces. And every time you ask it, it's going to probably give you a slightly different variation of this particular recommendation. Another thing that you can do is you can create rubrics with it. So I'm going to paste in the instructions. Again, let's see. Not this one. Did I paste in the right instructions? I don't think I did. Well, I got that weird thing. Anyway, I'm going to go on to the rubric. I'm going to ask it to. I'm going to ask it to create a rubric with three levels and three categories. Graduate level students for multicultural. So it's going to create two types of rubrics for me, a generic rubric for this task, and then it's going to create another rubric specifically for multicultural students. Now, what happens here sometimes is that there are two API calls. One create a rubric. The other one is criticize the rubric you made. So that is appropriate for multicultural students. And sometimes what it does is that criticizes this rubric instead of redesigning it. So there's still bugs there that I haven't fixed. But here's the generated rubric. We have three levels. We have three criteria for that set that I just put up there. And I don't know if you can see what it is or I wish I could just copy and paste it and put it into the chat, but I can't right now. But again, you have the link so you can play around with it and it will create a rubric. Once I that now that I have this, it will create a rubric for multicultural students. And I'm hoping that it will create one soon. Otherwise I'm going to skip it and go to. There we are. This one's much larger now, and I'm not really sure if I have to take a look at it in detail to see if this rubric is actually in line for what I want. The point is it will give you ideas As a starting point, it would be very, very wrong to copy and paste this and put it into your course outline, but it will give you some ideas on on how to do this. There's one other thing that I want to show you this is an inclusivity assistant. It takes your reading list and tells you how inclusive it is. So what I have here is a list of DOIs, just random DOIs I got from this article and I'm going to create generate report. And what's happening now is that is the AI is going to a library called CrossRef, which is both PubMed and Elsevier. And it's collecting data on these DOIs. It collects the author, the title, the journal, the year, the abstract, the funding info, subject categories, language and automatic score if it has. And it's going to save it in terms in a table in a JSON format that GPT3 just loves. And it will create a report on how inclusive it is on different categories such as gender balance. Here we are. So we have gender diversity, we have racial, ethnic diversity, geographical diversity, interdisciplinary diversity, representation of marginalized groups, temporal diversity, balance of established and emerging voices, balance of theoretical versus empirical work, range of methodologies. All that information is there from, from the data that it gets from Elsevier. So yeah, I'm just going to stop sharing now. So there are different ways AIs can help with EDI, both in lesson design and well, mostly in best in design. But I'm really hoping that this is just proof of concept. I'm really hoping that we can have a smarty in every school and every class so that teachers can create content very quickly and efficiently. Yeah. Thank you Mark.

Thank you Soroush and thank you for sharing Smarty. Making it available so it's available on the website as one of the suggested activities to take a look through it. And I know that every time I look at the application Saroosh, you're continually making improvements to it. I just used it for a course that I'm teaching this spring term and I had intentionally selected readings. I, you know, I tried to look at gender balance and where the authors came from and what the topics were and all of that. And the report I thought was pretty consistent with what I had attempted to do and it also pointed out some of the flaws. So I found that really helpful. So really appreciate that.

Thanks. Thank you Barb And I just want to shout out to everyone, so this isn't just me waking up one day and writing a program. This is many, many iterations of, of work and feedback from my colleagues at Worklands, particularly Barb Steam. We are here at this point and hoping that it will just get better. So if you do use it and find flaws or ways of improving it, please contact me I would be really, really happy to hear your feedback. Thank you.

Sorry.

I'm wondering if there's any plans for direct LMS integration or something that would just do this automatically as a matter of course. It would be good to see that, you know, we routinely assess for inclusivity in reading materials, for instance. Any thoughts on that? Or perhaps you haven't got that.

That would be a high level conversation with, with Brightspace. Not yet, no. Yeah, this is just experimental proof of concept level, level.

I'll call up John Baker and see what we can do. Sounds good.

Alex. I think that could be a, you know, a next step. I also like your idea around checking about integration with Curriculum Links. Like I think we just went through a, you know, curriculum review with our graduate programs and this would have been really a really interesting integration there as well.

Assessing the objectives within Curriculum links and then, you know, pushing it back into the system to ensure that curriculum mapping is properly done. That would be fantastic.

Yeah.

So, yes, lots of future projects for all of us to work on. That's great, great. Brenda, I'm going to turn it over to you.

I'm.

I'm wondering if you could share with us some of the questions that you have been using. I know you've had some great question prompts that you've been using to help colleagues, research teams consider accessibility in assessment in, in different ways.

Yeah, I'd be happy to share some of them. And a lot of it I think goes back to going back to the heart of what we're assessing and maybe challenging some of the embedded beliefs that come from our historical assessment practices. And I usually like to tell people, like if you want, the origin of the history of the multiple choice question comes from a reading test. So even, you know, almost 100 years later, it's still assessing reading ability, but we use it as a tool to somehow measure students recall of information or their ability to in other situations. So as we have AI coming in there, some of the ways that we can think about integrating it is really thinking about what are we trying to assess in our learning practices. And perhaps just like there have been lots of parallels to calculators, people still know how to add in their head, but it allowed us to perhaps have offload some cognition. Is this a new tool to help us offload some of the cognition and the executive function of writing by having a partner, someone who I can say help me with the scaffolding, help me figure out what would be the steps in terms of time management or planning. And I think that's an interesting way to start thinking about the tool as not something that is replacing a skill, but is helping to facilitate it and help us actually be more responsive to a very different looking in my context post secondary set of students. So students who are going to our postsecondaries today are not the same as 10 years ago, 20 or 30. If I go too far back, I know even myself wouldn't have been included in this environment, let alone as a woman. But as an individual with a disability, I think this is where AI poses a real opportunity is if it destabilizes our assessment practices because we are worried that students will cheat with my little bunny ears. It's also an opportunity for us to reflect on how we've not been responsive in the way in which we've designed our assessment activities. So if ChatGPT can answer your exam questions or your essay questions, much like Sarah highlighted earlier, is it a good question? What am I attempting to evaluate? What is the value add for the students? And as we move forward, I think this will give us some more flexibility in terms of AI to really allow students to play and explore other ways of demonstrating their knowing. Whether that be through can I design a good prompt? Can I use those prompts to create a new image that represents what I'm trying to get across? Does it open up ways for storytelling and all of those ideas? So really thinking about how can we make assessment more inclusive and if we need a catalyst, maybe AI and that fear and destabilization could be a good catalyst for us moving forward.

Thank you Brenda. I think related to your comments, Sarah has posted a link around cognitive offloading. Thank you Sarah for sharing the link in the chat as well. Taking a quick peek here, Sarah, I noticed you've also posted in the chat the tenant, so maybe I'll I'll turn it over to you and and ask you to share a little bit about either the tenants or I know you've been doing a lot of work globally, so maybe the global trends that you're seeing around AI and ethics and how do you see AI shaping the future of education on a global scale?

Oh man, that's a big question. I mean for sure this is a time of great discovery and also a lot of uncertainty. And in conversations I've been having with folks, been just kind of planting the idea that any children now who are 5 years old or younger, they'll never know school without artificial intelligence. And so to, you know, for teachers as they're thinking about their next school year, like this is the reality And I think so many educators and administrators are pretty unprepared for, for that reality yet. And a lot of us will be finding our way, sort of building the plane as we fly. Maybe we're building the AI as we, I don't know, learn how to use it. Not sure how all that unpacks, but kind of a key pie piece of all this around, you know, integrity and ethics and misconduct is this idea that we can, if we want to kind of build on Brenda's idea, on some of the cognitive offloading, we can relinquish some or all of our control to these apps. I can have an app do all of the writing for me that I'm relinquishing my control over my writing and my, you know, assessment completion to the AI. I can relinquish some of my control to the AI. Kind of like you were, you know, you talked about earlier your hybrid approach. You can engage with fact check checking, but ultimately what we don't relinquish is responsibility. So humans remain responsible for the outputs that we have. And it's one of the reason that the big scientific publishers, Nature and others have said that, tested that ChatGPT can't be an author. And I think, you know, some of the rationale behind that is because ultimately humans are held responsible for scientific and scholarly outputs. We're not yet at the point where we can hold an AI responsible because it's, it's, it's based on an algorithm, right? It's outputting something, but it's not yet sentient. I know that that's a contested idea. There are those out there who claim that AI is already sentient, but the general consensus still seems to be that sentience and responsibility, accountability remain uniquely human activities. I don't know if anybody's checked out the book yet, the non computable you. I'm just reading it now, really enjoying it. But it talks a lot about the things that AI can't do and a of it is focused around ethical issues and also relationality and human interaction. So yeah, where are we going with this? I think the future is still yet unknown, but worth paying attention to. You know, we have every indication that Microsoft Copilot and Google will have fully integrated AI capabilities with our everyday products, Microsoft Office and the Google suite of products here within probably the next 12 to 24 months. Months. And once we can sign into our computer in the morning and we'll have an AI app, summarize all the emails that came in overnight, draft our responses, build our to do Lists. Wow.

Maybe.

Maybe life just got a lot easier. And at the same time, there are ethical issues with that, because as an AI tool is doing all that, there is an element of surveillance behind it where the AI app knows what's in our email, how we're responding, who we're interacting with, with maybe that's already there, but it will become, I think, much more in the foreground. So there's still a lot to be unpacked and disentangled with all of this. I think. I think it's kind of exciting, but I still think that there's some cause to, you know, to just pay attention to some of the ethics behind it.

Thank you, Sarah. And did you want to say anything about the tenants? I feel like that up. I know, I know you've put it in the chat. And I'm just thinking, like, that's one kind of starting point that can be a way that, you know, that educators can start thinking about how we. How we navigate these. Yeah.

We talked about a few of them. You know, that human AI writing is likely to become fairly normal. You know, ultimately we still remain responsible for the output of, you know, our. Whatever we're submitting, whether it's a scientific paper, an assignment of a paper for our boss. I really don't think that human creativity is in any way threatened by AI, that our ability to imagine and create and inspire others, that remains boundless. And I think that's still a really, really important aspect of education and for teachers, you know, our ability to work with our students to encourage their creativity. We have great colleagues, colleagues who. Who specialize in creativity. So I'm excited to see where they take some of these technologies in their works. Probably the most contested of these tenants is one I worded quite poorly and take responsibility for that. And I said, well, language barriers would disappear. And I propose that one's first language, you know, might become, I think I said less important. And I didn't mean that in a sense of one's mother tongue becoming less important in, you know, an essential way, but rather that language barriers could be overcome. We already have apps on our phone. If you're traveling to a place where you don't speak the language, you can take a picture of something and it will translate it into your first language. So. And I confess that I was a child raised on the original Star Trek, so the idea of a universal translator was really seductive to me and continues to remain like, a great idea. Wouldn't it be wonderful if we got to the point where we could do that, but kind of circling back to some of those things that Alex and others brought up around the biases that are embedded in the language. Right. That language translation, I recognize, is a political act in itself. Interpretation is a political act. So it will be fraught with ethical issues for sure. But over time, how great would it be if we had something like a Star Trek universal translator? And then on kind of the other side of things, attribution will continue to remain important. And I can say that, that over the years of doing this academic integrity work, I've become really a lot less concerned with the minutiae of APA and really more concerned with the intentionality behind, like giving props to those that we learn from, whether it's through reading an article, listening to a story like the way that we show respect to those from whom we've learned. I think the intention behind it is ultimately more important than, you know, are we using sentence case capitalization or title case capitalization and recognizing that the style guides that we use are produced by organizations that profit from the sale of those guides. So that is a part of the commercialization and commodification of higher education. And so thinking about the ways that we can give acknowledgement can be more important than did I use a comma or period in the right way? I do think that our historical definitions of plagiarism are going to be disrupted. We saw some great examples of that in our chat tonight. In our conversation, cut and paste plagiarism is very likely going to be a thing of the past, if it's not already and some of you may have seen in the news. Soroush, I think posted it on Twitter the other day, that for companies such as Chegg, for example, with the stock price having plummeted quite recently, allegedly in part to the surprise of ChatGPT and other artificial intelligence, in my circles, folks are asking and assuming actually that technologies such as ChatGPT will make contract cheating a thing of the past as well. So I. Yeah, plagiari definitions of plagiarism I think are going to have to change because what we'll be dealing with will be completely new and different.

Thank you, Sarah. And you'll notice there's some, some further comments in the chat you might like to take a look at. I know we're getting really close to closing the session, so I, I wouldn't mind, Saroosh, if, if there's something you'd like to say, perhaps about how you're offering guidance and I know you've created some guiding principles that Might be something you'd like to share.

Yeah, thanks Barbara. I'm going to put, in the interest of time, I'm going to put two links into the chat which I think are two documents that I think are really really important when it comes to equity and ethics and responsible use of AI. I've been really looking at these two documents very closely and I have tried to pull out of these documents terms of use for, for our students in higher education to be used on course outlines. And I kind of put everything into this little paper here. I'm going to just post it onto chat. So just keep in mind that these are just my thinking on what I think AI terms of use could be in the classroom. It was derived from these two documents among others that I have read. But at a very, very. I'm just going to do this very quickly play. So at a very high level it covers basic tenants Respect human rights and dignity. So when you using AI tools make sure to respect human values like privacy, dignity and human rights. Don't use AI to identify your classmates emotions online and then ridicule them or belittle them them. Ensure fairness and non discrimination. Choose AI systems that are unbiased and promote equal opportunities. Don't use one particular AI to make an argument that you know that it's going to be biased or it is biased. Ensure fairness and non discrimination. Choose AI systems that. Let's see, I talked about that. So prioritize privacy and data privacy protection. So handle personal data very responsibly and use AI that take that privacy very seriously. Promote transparency and explainability. So that's. That means understand AI systems use in course work and be ready to explain exactly how the AI contributed to your academic work. Commit to safety and security. So use AI's safety without compromising security, security or reliability. Basically means don't let. Don't give AI login credentials and let it do something for you. Maintain human control over technology. So use AI as a support tool keeping human agency at the center of educational experience. Don't over rely on it. Demonstrate professional responsibility. So use AI ethically considering the long term effects of collaborating with others to address ethical concerns. So for example don't use AI to manipulate data or misrepresent research results and foster and promote human values. Use AIs in ways that align with societal values and contribute to human well being. So don't use it to create fake news or incite hostility at all. So these are some of the ideas I have have. I'm just going to stop here. I'M going to put this little chart into the chat and maybe it could be this conversation service sometime.

Great. Thank you so much. And you'll notice on the website that we've included some links for you. We can add some of the ones that were added in the chat tonight. There's also some activities that you can engage in if you like. So, for example, the McLuhan's Tetra, if you want to take a look and just use it as a reflective kind of tool to be able to think about the, you know, the possibilities, as I said, or the challenges that have been presented through this, you know, emerging and continuing to emerge technology, then that might be a way to, to further engage in thinking about the topics that we've shared with you and to continue discussing them as, you know, in Discord or in other channels that you're using. So I would like to thank the panelists for joining us this evening as well as all of you. And for all of the I couldn't keep up with all of the chat going on, so read some of the messages, but not all of them, but appreciate it. Appreciated that. And if you want to include a little message in the chat as your way of ending the session, and that would be welcome as well. So thank you again.