WS #232 Innovative Approaches to Teaching AI Fairness & Governance

16 Dec 2024 11:30h - 12:30h

WS #232 Innovative Approaches to Teaching AI Fairness & Governance

Session at a Glance

Summary

This workshop focused on innovative approaches to teaching AI fairness and governance, emphasizing the use of serious games and project-based learning. The speakers discussed the importance of making AI education more inclusive and accessible across diverse educational systems globally.

Tayma Abdalhadi highlighted the need for interdisciplinary approaches to AI education and the importance of empowering users through effective feedback loops. Ayaz Karimov introduced the concept of serious games, explaining their effectiveness in teaching complex subjects like AI. He demonstrated this with an example of a cybersecurity game called “Duck Code.”

Melissa El Feghali discussed the benefits of project-based learning in non-formal educational settings, emphasizing its flexibility and potential for community engagement. The speakers addressed challenges in implementing these methods, including issues of access to resources and expertise.

The discussion touched on the importance of cultural adaptability in AI education methods and strategies for measuring the effectiveness of interactive tools. The speakers also addressed concerns about screen time and the responsible use of AI tools like ChatGPT in educational settings.

Key recommendations included focusing on grassroots efforts, creating flexible learning frameworks, and promoting AI literacy. The speakers emphasized the need for contextualization in educational games and the importance of teaching critical thinking skills when using AI tools.

The workshop concluded with a call for educators to adapt their teaching methods to incorporate AI tools effectively, encouraging students to use these technologies as aids for creativity and critical thinking rather than as substitutes for original work.

Keypoints

Major discussion points:

– Using games and gamification to teach AI fairness and governance concepts

– The importance of project-based learning and non-formal education settings for teaching AI topics

– Strategies for making AI education culturally adaptable and accessible globally

– Challenges in implementing AI education initiatives, including access to resources and expertise

– Concerns about screen time and responsible use of AI tools like ChatGPT by students

Overall purpose:

The goal of this discussion was to explore innovative approaches for teaching AI fairness and governance, with a focus on interactive and engaging methods like serious games and project-based learning that can be implemented in diverse educational contexts.

Speakers

– Raneem Zaitoun, Workshop moderator

– Tayma Abdalhadi, Research Analyst on Human-Centered Technology, UN Regional Youth Advisory Group

– Ayaz Karimov, Researcher in Gamification and AI Education, Head of Product, Swiss Cyber Institute

– Melissa El Feghali, Youth Representative, World Organization of the Scout Movement (WOSM)

Full session report

Innovative Approaches to Teaching AI Fairness and Governance

This hybrid workshop, featuring both in-person and online participants, explored innovative methods for teaching AI fairness and governance. The discussion focused on interactive and engaging approaches that can be implemented across diverse educational contexts globally, featuring insights from experts in human-centered technology, gamification, and youth education.

Speakers and Their Presentations

1. Ayaz Karimov – Researcher in gamification and AI education, Head of Product at Swiss Cyber Institute

Karimov introduced the concept of serious games as an effective tool for teaching complex subjects like AI. He distinguished serious games from regular games, explaining that while regular games are played for entertainment, serious games have specific educational goals. Karimov demonstrated this approach with an example of a cybersecurity game called “Duck Code,” which was played at the Global Cyber Conference in Switzerland. He highlighted key elements of effective educational games, including narrative, progressive difficulty, realistic scenarios, and clear objectives.

2. Tayma Abdalhadi – Research Analyst on Human-Centered Technology, UN Regional Youth Advisory Group

Abdalhadi emphasized the importance of an interdisciplinary approach to AI education, involving both technical and non-technical perspectives to provide a comprehensive understanding of AI concepts. She discussed the dynamic relationship between users and AI technology, highlighting issues of algorithmic bias and the importance of empowering users through effective feedback loops. Abdalhadi stressed the need for cultural adaptability in AI education methods, adapting learning materials to local realities and cultural contexts.

3. Melissa El Feghali – Youth representative at the World’s Organization for the Scouts Movement

El Feghali advocated for project-based learning in non-formal settings, allowing for more flexible and inclusive learning environments. She emphasized how this approach breaks down barriers of participation for youth, fosters collaborative learning, and enables real-world customization. El Feghali noted that the impact of project-based learning extends beyond the immediate activity, creating a “ripple effect” that continues to influence learners and their communities.

Key Themes and Approaches

1. Interdisciplinary and Inclusive Education

2. Serious Games and Gamification

3. Project-Based Learning

4. Cultural Adaptability and Contextualisation

Challenges and Concerns (from Q&A)

1. Screen Time Management: Karimov raised concerns about managing screen time, particularly for younger learners. He introduced the concept of ‘cyber disease’ and emphasized the importance of limiting screen time. El Feghali suggested that many educational games could be designed without requiring screen time at all.

2. Access to Resources and Expertise: El Feghali highlighted the lack of access to digital tools and expertise as a significant barrier to AI education in many contexts.

3. Responsible Use of AI Tools: The speakers addressed challenges posed by AI tools like ChatGPT in educational settings, particularly concerns about cheating. Abdalhadi emphasized the need to teach critical thinking skills when using AI tools, encouraging students to compare information from multiple sources and understand that AI outputs are not always entirely accurate.

4. Contextualizing Project-Based Learning: El Feghali discussed the importance of adapting project-based learning to local contexts and resources, emphasizing flexibility and creativity in implementation.

Recommendations and Future Directions

1. Focus on Grassroots Efforts: Emphasize local initiatives and community engagement in AI education.

2. Create Flexible Learning Frameworks: Develop adaptable educational approaches that can be tailored to different contexts and learning styles.

3. Promote AI Literacy: Karimov stressed the importance of understanding AI concepts, including technical aspects like hallucination in language models.

4. Adapt Educational Goals: Abdalhadi suggested reframing educational objectives to focus on critical thinking and creativity rather than just content production when using AI tools.

5. Implement Feedback Loops: Ensure that users of AI technology have opportunities to provide input and shape the development of AI systems.

6. Encourage Collaborative Learning: El Feghali emphasized the value of peer-to-peer learning and community engagement in non-formal educational settings.

Conclusion

The workshop concluded with a call for educators to adapt their teaching methods to incorporate AI tools effectively. The speakers highlighted the potential of serious games, project-based learning, and culturally adaptive approaches in making AI education more accessible and engaging. They emphasized the need for ongoing research into the effectiveness of these methods and the importance of developing AI literacy across different age groups and contexts.

The discussion revealed diverse perspectives on implementing innovative approaches to AI education, with speakers offering complementary strategies to address the challenges of teaching AI fairness and governance in a rapidly evolving technological landscape. At the end of the workshop, participants were provided with a digital handbook and additional resources to support their efforts in AI education.

Session Transcript

Raneem Zaitoun: Hello. Thank you so much for joining us in person and online and welcome to the workshop. So workshop 232 and innovative approaches to AI teaching and fairness. A little bit more about our session today. So we will be exploring innovative methods to teach AI fairness and governance and this workshop will combine some expert insights, interactive sessions, example games, as well as policy discussions to foster inclusive and impactful discussions around education. So if you please move on to the next slide. Perfect. So let me introduce to you our speakers for today. We have over here Ayaz Karimov. He is a researcher in gamification and AI education and also a head of product at Swiss Cyber Institute. We also have our online speaker Tayma Abdelhadi. She’s a research analyst with a focus on human-centered technology. She’s a United Nations Regional Youth Board member as well and over here we also have Farhadi. She is a global youth representative at the World’s Organization for the Scouts Movement. All right. Next slide please. So for our workshop agenda today we will be delving into serious games and how we can use serious games for AI education and teaching AI fairness. We will then discuss policy questions on inclusivity, frameworks and the like. We’ll delve into an example game session for you and we’ll be talking about project-based training for practical learning as well. we’ll finish it off with a Q&A session where you can ask questions to any of our speakers, and we will also be asking questions to the speakers itself. And so without further ado, I think we are going to get started with the objectives of this session. If possible, could you please move to the next slide? Next slide, please. Yes, awesome. So some of the objectives for this session will be exploring AI fairness principles, discussing some frameworks around governance and AI, and introducing innovative teaching methodologies to enhance understanding. As this is a hybrid setup, we are looking to have our online participants ask questions through the Zoom itself and interact online, and all of the in-person attendees, please save your questions for later during the Q&A session, and we’ll have an interactive segment later on. Can we please move on to the next slide? And the next one, please. All right, perfect. So I’m going to get started with our online speaker, Tayma Abdelhadi. Tayma, if you could get started for us. So I’ll pass it off to her. Hello, everyone. Hello, Tayma. We can all hear you. Go ahead.

Tayma Abdalhadi: Thank you so much for this introduction, and thank you so much for this very, very important topic and session today. I’ll be the introduction for this session, and I’ll start from the fairness topic. And I think this is important because we have been pretty much inclusive in discussing policies and discussing how can we make this more fit with the human rights sector, but we are not inclusive when it comes to technology. Oftentimes, we keep the information and discussions regarding to how we shape the AI to the technical people, and I think that’s very, very wrong because it’s pretty much an interdisciplinary field right now. What I mean that it’s a dynamic relationship between the users of this technology and the technology itself. When we use the AI model, we shape it by the information we give it, but it also shapes our understanding of the topics that we give it and ask it about. That means it’s no longer just an input-output aesthetic output situation. It’s a dynamic situation, and if we do not empower the users by knowledge and by feedback loops, we will have a pretty skewed AI model, or we will have frustrated users. This leads to multiple problems. The first one is when you’re focusing on algorithms themselves, we can have bias. This bias can be eliminated in two ways. The first way is by the makers themselves who are our students, our future makers of technology. If we manage to install those concepts of fairness and how can we think about the human that we’re making the technology to as the code is being automated and we’re allowed more time and space to think critically and imagine scenarios, then we can mitigate or at some point create safeguard methodologies to protect the users against algorithmic bias. The second part is the users themselves. If we empower users through an effective feedback loop, that means we can officially install and mitigate further damage by the bias that’s held. happening. I’m totally against a concept that was installed traditionally, which is if the product is wrong, then we retrieve it, we fix it, and then we send it back. But right now, we don’t have that luxury because you simply don’t have any competitive context as a user. If you use, for example, ChatGPT and it gives you an answer, then you assume that’s the full truth answer, unless you take it and then you spread it on Reddit, like, for example, the David Mayer case, where some users raise concerns that we do not get any information if we ask it about certain names. It could be a privacy tactic. It could be a censorship tactic. But we don’t have enough transparency from the AI model that tells us what is exactly happening. And the most dangerous thing is when it only gives us a little bit of information and we think that’s all information available. One example that this could be dangerous in is when a political, when you ask it about a political figure and it gives you all the great, nice stuff that it did, but then it does not give you the part where you might think badly of them. This can be used. This can abuse the Privacy Act, for example, in the EU. And it can allow certain information to go through AI models and certain information not. While the user thinks this is the full truth and this is an unbiased opinion. This is also in the perfect cases. But we don’t also talk about the unintended cases that AI might be used in. One of my colleagues is in Africa and they say that sometimes when the teachers in elementary school have internet, they try to get pictures online to show their students about animals, about other countries. And the only truth they know is that static pictures. So imagine those communities being a target to a misinformation campaign where you have AI generated images. How can you explain? this audience that only knew the truth through static images because they don’t have enough access to internet, that there is something called AI generation and these images need to be fact-checked. I believe this is very, very important. I think role scenario playing and serious games and education is not just for computer science students or people in that disciplinary, but it’s also a part, it should be a part of the policies of companies who are deploying this technology as a way to humanize and to imagine further scenarios of how the technology is being used in this speedy, extremely fast-moving world. We need to think for two minutes and say, okay, if I deploy this tomorrow, what might the impact be? And who do I need to measure and imagine that impact? And the second question is more important. And I believe this is one of the key messages that we’re here to deploy. And this is part of conferences like this, where we bring diverse people and have messages discussed and see other aspects too.

Raneem Zaitoun: All right. Thank you so much, Tayma. If we could go back one slide, please. All right. So now I’ll pass it off to Ayaz to speak a little bit more about the use of serious games for teaching AI fairness. Yes.

Ayaz Karimov: Yeah. I can hear myself. So it means actually you can also hear me. Today, I will talk a little bit about the serious games and I will also show you one game that I made and we played last month in Switzerland at Global Cyber Conference. So it could be good examples of actually how we use the game there to teach the cyber security or cyber literacy. But before I start talking about everything, let’s talk about the games and the serious games like that. Maybe it’s the first time even you heard about this term serious games. The main difference between the game and the serious game is that actually when you play the game, you just play for the sake of, like, getting entertained, just to achieve something. But in the case of the serious games, you totally have totally different ultimate goal. For example, let’s say, actually, if your game is about health, let’s say, then your ultimate goal is to get healthier. If the game is about education, then your ultimate goal is to learn something or to teach something in a very, very good way. So you don’t care, actually, if you get entertained or not, but your first goal is totally different thing. But of course, then it comes to entertainment, the other stuff. So the main difference between the serious games and the games is that actually you have the one goal, and beyond this goal, you also have some kind of things to achieve, let’s say. I also did the research about the serious games, whether they really help the learner to achieve something or not. But I also put another research that was carried by other co-researchers about the serious games and their effectiveness. And the research shows that actually when you use the games, it generally positively impacts the motivation, engagement, and academic outcome. So in the general case, actually, you will have very, very good results from the games. But having said this, I also want to say that actually not all games are good for everyone. So hereby, we need to highlight the importance of personalized learning. But as I mentioned, in most of the cases, the serious games are considered as a very, very efficient tool to teach hard topics, something like AI or, let’s say, the STEM subjects. And why these games are good, another perspective is that most of the games are simulation-based. So even if you play the board game or, let’s say, if you play the other strategy games, in most of the cases, you get the simulation from the real life. So while you play the game, maybe your brain doesn’t understand, actually, what’s happening there. But actually, you learn something there which can be implemented in the daily life. in the real life as well. So that’s very, very strong tool actually nowadays. And I also put in the last bullet point, I put some types of the game, but today I will particularly talk about the puzzle game, which is very, very good way of this puzzle game is the escape room game. And what happens there is that actually you have the challenge in the game and you try to do some kind of actions. You try to solve this challenge. And once you solve this, then you move to another level and another level and you just try to escape the room, let’s say. If you can go to the next slide, please.

Raneem Zaitoun: And next one, please. So we cannot see the slides, like if you’re sharing them because we cannot see them. Sorry, everyone, a bit of a technical issue. All right, while we work on that, I could get started.

Ayaz Karimov: I will just briefly talk about this game. I will stand up because I also put some games from this games that I prepared with one game designer from Portugal. And the game’s name is the Duck Code. And the idea is that actually, if you can go to the next slide, please. So the idea is that actually we have a hacker and the hacker did the hacking in the four cities previously in Europe. So the story starts with like that. And then it’s the one hotel in Zurich. And the idea is that the fifth attack was going to happen or is going to happen in this hotel room. And the players are put in this hotel room and they try to make sure that actually they find some kind of clues so that they can stop this hacking. So the narration of the narrative of the game works like this. And it’s pretty much simple clues. So we didn’t use actually anything hard because the idea was that actually, if you put here like some kind of coding challenge or something like this, not everyone is going to. solve this. So we had really, really simple clues, which also helped them to learn some kind of cyber literacy, but at the same time, really get entertained. Can we go to the next slide, please? Yes. So again, we had the four hacks happened, and the next one is going to happen in Zurich. And our hacker left some of his belongings in this room. And in this, I will show you some of these items in the room so that you can understand what kind of belongings we are talking about, and why even we use these belongings in the game, because I guess that’s a very important point to highlight. If you go to the next slide, please. So now in the next slide, I will talk about six different things to you. So if you want to use the game, or if you want to create your own game, then you can just take a note, or I don’t know, just keep in your mind that actually, those are the most efficient, or those are the main techniques that actually you need to implement. So here, our question is that what you can do as a professional to make sure that the game is engaging, and it really teaches something. If you can go to the next slide, please. So the first one is the narrative. When I start introducing the game, if you remember, I started directly talking about some kind of background information, right? Like the hacking happened in some countries, it came to the Zurich, and actually, we had the one game master. Basically, this is the person who was leading the game. The game master was pretending to be the detective in the game. So he was in the same room with our players, and he was showing some kind of things, and he was just trying to pretend that actually he is also trying to solve this with them. So the first rule is that actually, you really need to have the good narration in the game. If you don’t have the narration, it generally fails, actually. Can we go to the next slide, please? This progressive difficulty, what does it mean, actually? In the first clue, let’s say that in your game, if you have the three or four clues, you always should start with the basic one. And if you use the basic, basic, basic, basic, then most of the time, actually, your game fails again. So the idea here in this game techniques is that you always have the level one, let’s say difficulty one. In the level two, you have at least a little bit more difficulty. If you don’t have the more difficulty, then it’s also another problem. And in our case, we also implement the same thing. And here, maybe it’s also good to mention, especially you don’t have to have so many challenges in one game. We had only two challenge, actually. We just had the one basic and one very, very hard one. Can we go to the next one, please? And the realistic scenario. So here, we bought a pot from the shopping. And what we did is actually we broke this pot in the room. And why did we do this is actually, normally, if you run away from there, highly possible is that actually you crash something and something fall down, right? And in the game, we did these things a lot. Like we just threw away some books, some glasses fall down, and that there was some juice on the floor. These kind of things, actually. When you have these really realistic things that actually happen, or some kind of actions in the game, the most players actually really, really like it. You can also integrate this kind of realistic things in the game as well. For example, in our case, we were using the pieces of the pot here, just pieces of the pot to create some kind of challenge for the players. So that’s why we didn’t only create this realistic scenario, but also use them as part of the challenge. Can we go to the next slide, please? The clear objective. So when you play the game, at least I guess it’s the core of all the learning activities, not only the serious games, but it’s pretty much important in the games as well. So basically, when you have the game, you cannot tell the players that you will win when you escape the room, right? It’s not that pretty much objective. You have to be clear. that actually what are the exact things that actually they need to do. For example in our narration we started by saying that you will have the two clues and that one the first part is there and the second part is there so they had the clear goals or clear objectives actually what they are going to do. By the way in this game just since this is in the picture I also want to say is actually we use exact documents that we try to find from the public resources as a crime office so we really pretend like it’s the real document from the crime office that’s why it has too much detail there. Can we go to the next one please? Yes and now actually my colleague Melissa is going to talk about the project.

Raneem Zaitoun: Yes so now we’ll pass it off to Melissa to talk about project-based training for practical learning.

Melissa El Feghali: Hi everyone so we’ve seen through the interactive example that Ayaz gave us how we can transform very abstract ideas such as AI fairness into tangible and engaging ones but the idea is how do we scale these approaches and how do we use project-based learning how do we use it to transform these experiences into sustainable and teaching methods and impactful methods that we can use in that are non-formal such as scout groups, community spaces, youth movements etc. So I’m going to talk about three things the first answer is why use gamification for project-based learning so gamification that uses games challenges and simulation is a natural fit for project-based learning but why because first of all it encourages active problem-solving and critical thinking but also it creates a very flexible environment to learn. So it makes learning accessible, because usually we’re used to traditional curricula where we have to follow it. But in these examples, we have the flexibility to experiment around in the places that are doing these games. And also, it builds engagement through clear objective, realistic scenarios, and the progressive difficulty that Ayaz talked about previously. The second question that you might ask is, why use these settings, and why do these settings work? So non-formal educational environments, such as scout groups, youth clubs, and community spaces are ideal for project-based learning, because they break down the barriers of participation for youth. So we have more flexibility to experiment, as I said, and we don’t have the barriers of who has access in terms of financials to access these places or not. And it fosters collaborative learning, where people work in groups, and they learn about empathy, and they build collectively ownership over the solutions that they come up with. And finally, it enables real-world customization. So youth can adapt challenges to local contexts. So they can adapt the game, for example, to local AI fairness issues, such as accessibility to schools or the technological resources that are available or not. The third thing I want to talk about is impact. So how do we scale the impact? Project-based learning doesn’t just stop when the activity ends. It creates a kind of ripple effect that goes on. So if, for example, someone in a scout group creates an AI challenge game about AI fairness, they can take this game to another scout group, to another school, to another youth club. And so from one idea, you can generate multiple ideas, and it creates this ripple effect, because participants gain ownership of the game that they’ve created. it can be adapted to multiple communities. So it’s not just like a fixed game that I’ve created and it only works within their context. So I just want to say that project-based learning supported by gamification empowers communities to democratize access to AI fairness and AI fairness education. So by integrating clear objectives, by creating realistic scenarios, by doing this progressive difficulty, a way of learning stuff, we can bridge the global concepts that are sometimes very complex and that we’re not able to break down, such as AI fairness, to local context and lived realities and we can make learning more inclusive, engaging and scalable. So one last thought to leave you with is think about the communities that you live in, think about the youth that are around you within these community spaces and how can you use project-based learning methods to introduce them to topics such as AI fairness and ensure that the message scales beyond just the session that you’re giving or the room that you’re in. Thank you.

Raneem Zaitoun: Okay, great. Can we move on to the next slide, please? Thank you, Melissa. All right, so now we’re launching into a Q&A panel discussion. We’ll start off with a couple of questions we have from our end then I’ll open up the floor to any questions that the audience may have. So I will start off with our online speaker, Tayma. So Tayma, my question for you is how can we ensure that the methods used to teach AI fairness and governance are culturally adaptable and accessible across diverse education systems globally?

Tayma Abdalhadi: I think this has two main factors that we need to take in mind. The first one is grassroots effort. It’s really damaging that technology comes from top to bottom when we go from a global scale to local communities because no matter how we try to understand the context we can never get it as the person who’s living there. And luckily, because the technology is such spreading worldwide, we can always find individuals within those communities who understand the local reality but also has the enough qualifications or at least the ability to connect with those global entities and then get them to speak to the local entities’ circumstances. Whether it was internet access, any cultural difference, it could be the simple story I told you about in classrooms where you don’t have much access to internet or materials. And if reached to the right person, it could impact a change in the technology itself. The second thing, as I always say, it’s the feedback loop. We can’t always say that when we put this little button that says feedback or support, that it’s enough motive or it’s suitable for everyone to report through it. We need to go and see what actually motivates people to make an impact and to deliver their voices. And most importantly, how can we make sure that those voices that have been delivered are actually getting the counterfeedback they deserve? So if I tell one company that their videos do not have enough markdown and they look too realistic and could be used in misinformation, and then I don’t see any impact, it will disencourage me to provide feedback again. And that will create frustrated users from one end, but also a broken loop of understanding and feedback from the other. So it’s mainly grassroot work. It’s empowering individuals and making their voices heard by updates, by working within the community to provide solutions that work for them.

Raneem Zaitoun: So I will now pass off the next question for you. What strategies would you recommend for measuring the effectiveness of interactive tools such as simulations or serious games and teaching complex concepts like algorithmic fairness?

Ayaz Karimov: I guess we already discussed about how games are good for this teaching, especially hard subjects. There has been other research actually why people are using the games. And actually one of the reasons actually, for example, if you want to teach biology or physics, it’s very, very hard subject to teach. So that’s why they were just trying to find some ways, and one of these ways is serious games. But I guess the question here maybe we need to ask is that what type of game I need to use? Because, for example, if I am a teacher, the first question that comes to my mind is that actually there are dozens of games, like the board games, simulation games, strategy games, like that. Which one to use in my classroom? And actually, the best way is to do this. Actually, you do some test iterations that actually one day you start with the board game. In the second day, you play the strategy game. In the third game, you play the puzzle game. And you just try to see actually which one is the best for your entire classroom, because I guess that was the best strategy to do. And that was the one thing actually we did with one biology teacher in our research. So I would say actually testing different games and trying to see actually which one is much more impactful, at least from the eyes of the teacher.

Raneem Zaitoun: Perfect. Thank you so much. Our next question goes to Melissa. So, Melissa, my question for you. What are the key challenges youth groups may face in implementing project-based training, learning for AI governance? And how can policymakers support these initiatives?

Melissa El Feghali: Okay. I’m going to talk about one challenge that I think is essential. It affects a lot of other things, and it’s access. It’s lack of access. Access in terms of resources, but also in terms of expertise. So a lot of youth movements lack the expertise in terms of mentorship, but also in terms of the technology that is being used or the tools that are provided to them. So three things that I think are very important to policymakers to take into consideration. The first thing is, of course, investing in infrastructure to have better access to digital tools, and also to digital literacy programs. The second thing is to create partnerships. Create partnerships not just between governments and private sectors, but also between the civil society sector. And in that way, there’s bigger room to share expertise from both sides. Third thing that is the most important is to support the flexible learning networks or frameworks that are available out there. So, when you give flexibility to the participants or to the people creating these initiatives, you have more room to experiment, to fail and then to build upon it. And this is a very important concept when it comes to project-based learning and the whole process that is used. So, if these three conditions are met, then for sure the challenges would be less challenging, let’s say, for the people who are trying to do this. And youth will be empowered in that way to have to not just learn about this concept but also be part of the solutions that are being created and have it be more flexible in terms of contextualization and the communities that they live in and that way it’s more relatable to them. Thank you.

Raneem Zaitoun: All right. Thank you so much for that, Melissa. Now, I will open up the floor to the audience if anyone has any questions to ask our wonderful speakers. Online as well. So, both online and on-site. Gulcan, if you have any questions through the Zoom Q&A, please let us know and we’d be happy to answer any of the questions as well. All right. Go ahead. All right. I’ll come up to you and give you the mic. Sorry. Excuse me.

Audience: Thank you. Thank you very much. Thank you very much for the speakers, for the great information. I have just only one question, which is related to the screen time and how we can control, knowing that these tools is really very beneficial and with a very good intention to improve and develop, how we can manage the screen time and maybe the impact of the screen time or the access of screen time. Thank you.

Raneem Zaitoun: Should I? Yes.

Ayaz Karimov: if whoever wants to go. Yes, very, very, very good question. Actually, it has a name in the academia as well. It’s called like the cyber disease like that. It’s not only when you watch the screen too much or when you use the VR or AR, your brain like the loses itself. And if you like to watch your phone too much, your head is going to ache too much. And that’s why it’s called like the cyber disease. That’s very good point. I know some information is actually what could help for this, but actually it is still very, very huge problems, particularly for this immersive technologies, because everyone is talking about this technologies, but when it comes to using them, actually, it’s very challenging because not anyone can use this for a longer time. So normally it depends on the learner’s age or a little bit more detailed, of course, is needed to say something exactly. But normally, at least from the one of the research that actually I did, if the children’s age is between the eight to 12, then the normal playing time for the educational games should be maximum 40 minutes. So it’s considered like the one lesson per day, it’s enough. Normally one lesson is approximately 40 minutes in the countries that actually I live in. So it’s 40 minutes is acceptable. But of course, if you get tired, I know that actually you get to play with more fun and there are also other dimensions as well. For example, are you using this time only to play some games? Because also some games are designed in a ways that actually it doesn’t impact you that much. So actually you don’t have to really make your eyes tired. So it really depends on the game. But most of the time, the shortest answer is 40 minutes if the children’s age is up to 12.

Melissa El Feghali: I can add to that on what Ayaz said. It really depends on the game. that you’re aiming for. A lot of the games that can be created don’t even need to use screen or have screen time. And this is exactly one of the points we talked about the flexibility of these kinds of settings. You can do a whole game about AI fairness and education and outdoor settings where you wouldn’t even need any screen or screen time and you can relay the message in a creative way. So I think it really depends on the game, how you want to design it, what’s the context you’re in and what are the resources provided as well.

Raneem Zaitoun: Thank you. Tayma, is there anything to add?

Tayma Abdalhadi: I would echo what Melissa said on the flexibility of hybrid games, especially for children. And I think that’s a very important topic because oftentimes we focus on the digital solution staying in a digital world where we miss the core idea that it depends on logic. And at the end of the day, if you want to humanize something, you don’t need a screen for you to tell you how to humanize it. You just imagine, you use your imagination, use relevant cases and scenarios and that does not necessarily need a screen for you to do.

Raneem Zaitoun: All right. Perfect. Thank you, everyone. Do we have any other questions from the audience? I’m standing up in case I need to hand the mic to anybody. Any other questions? All right. Perfect.

Audience: Yeah. So I think one question or query that I have is when it comes to contextualizing project-based learning or even engaging learning or non-form education, many times what happens is that if you do not bring in or the elements, trying to bring all those elements as it is, like bringing the originality, when you contextualize it, that tends to really degrade the level of engagement or the brightness of that product. So have you any examples of any contextualized project-based learnings or initiatives that really has depicted that originality or has not even degraded with what it is supposed to be? So we’d love to know that. My question is, recently there have been a number of parents and teachers complaining about children using ChatGPT, whether to learn or to do assignments. I’d like to hear your thoughts on that. Do you have any tips on how? how to teach your and use ChatGPT in your own context. Thank you.

Raneem Zaitoun: All right, so I guess if we could tackle maybe each speaker asked answers one question, I think it’s most efficient.

Melissa El Feghali: I can answer the first question. So I have a lot of examples we can even discuss after the session if you’re interested. It really depends on the thing to contextualize. So for example, I was part of a training that teaches about peace education and peace is a very broad subject that can be like perceived in a different way in different contexts. And so this program is done around the world in different schools and in different communities. The idea is you keep the same concept. You can even keep the same game, but you have to change the instructions within the game. So for example, we used to do a game about human rights. Okay. And we have to show the difference between visible and invisible violence, and we show them real life examples. So in every community or country that we go to, we change these examples of scenarios and get scenarios from within the same community or country so that the people so that the children can relate to the stories that they reading. So instead of just saying that bullying happened in a school, we give a specific example about a specific type of thing that is usually mentioned in bullying within that community. So that’s a very small example, but it doesn’t devalue the concept or the idea you’re trying to portray. It makes it even more relatable. So that’s the idea about contextualization. You can have a set of games that you can take around, but you can’t keep it exactly the same. So that’s just a very short example.

Ayaz Karimov: Right. Yeah. And I will take the second question, which was about this. Let’s start from one thing. If you didn’t know, already AI, at least the mother of this boy is saying that. Yeah. Yeah. Yeah. Killed his son. So if you didn’t read the news like in the beginning of the November one one guy did the suicide and that there is still in Investigation, but actually there is very high possibilities that actually it was because of the AI because AI basically was misleading and Again, we can discuss it actually how it happened and why it happened and it happened in the characterized dot AI if you don’t if you know the tool that was that so it’s not the charge of it him. I guess they are even using for other stuff and actually it’s not even the children like that I know pretty much many people that actually they treat you with as a dear Psychologists, which shouldn’t be the case so I guess here the main important thing is that actually this AI literacy come to the stage because when what’s the first education we get in our lives is that in the school at least most of us learn how to read and write and that’s called literacy and then you become a little bit older and older you someone teaches you how to use the computer and that’s called digital. They teach you though how to be safe in the Internet then you learn a little bit cyber literacy and I guess now it’s the time to highlight the importance of that literacy. Actually, many governments are doing some stuff about this, I know that the Netherlands and also in the United Arab Emirates. They have the huge initiative about this a elitist and a particular prompt engineering So that you don’t get to believe that what the AI always tells you but you also have the enough knowledge to understand what’s wrong and what’s not. But it doesn’t and it’s the same thing with the LLM’s this AI tools. They give you the something and they pretends that actually they are telling the truth, but they don’t tell the truth they just hallucinate because it’s just the element they just try to give you some random answer. So the AI literacy basically helps you to not to get to these hallucinations and to make sure that actually you use these things efficiently. So my short answer would be really focus on this AI literacy. And I guess it’s not only for the children. At least I know that half of my friends or people around me, they don’t have the enough AI literacy, I guess, to be able to use these tools.

Raneem Zaitoun: All right. Thank you. Tayma, is there anything to add for any of the questions?

Tayma Abdalhadi: Yes. I want to add something on the second question because I’ve been working a bit with Child Online Safety. And this has been a reoccurring question. And I think this has to do with how we’re approaching AI tools, and especially generative text tools. The first one is, it’s very, very important to make sure that the children know that this is not reality, or at least it’s not the full reality. And this is what I was talking about, about not having comparative contexts. Even us when we’re dealing with chat GPT, when we’re on our own in the chat, we think this is the full truth. And we don’t have any way to compare that or to verify it. So for children, it’s important to be like, okay, this is the answer from one AI model. How about we try other AI models? Or how about we also try humans and ask the teacher or the mother or the sibling? And that way, they are built with a comparative system that they can always verify and check whatever is coming out of the internet, really, not just AI models. This was also a problem with Facebook before AI models came about. It just got upgraded with LLMs. The second part is, what is the goal of the homework? And I think this is a critical problem for all educators. If the goal is just to memorize or just to achieve an essay, then you’re failing as an educator to actually adapt with the technology. And the important thing is that you’re trying to teach your child how to have critical thinking using the LLMs. So instead of being like, okay, instead of writing your homework, how about you ask it what subjects you can write about? And then also make sure they understand that whatever is coming out of this LLM should not be satisfying for them to present on their behalf. That they have this energy and they have this creativity and they have this style in writing that no AI language model can replicate. And no matter how perfect, there’s no such thing as perfect or this perfect answer that we can get from the AI models. It’s the answer that you actually generate using your imagination, your experiences, whatever language that you have. And this should be rewarded, not just the grammatically correct, the perfect language answer. And I think this falls heavily on the educator’s side.

Raneem Zaitoun: All right. Thanks so much, Tayma. Okay. Is there any other questions from the audience? Okay. So I think we are going to wrap it up. Can I please go on to the next slide? All right, everyone. Thank you so much for your time and your attention for this workshop. We have provided actually a digital handbook that you can access. If you could please go on to the next slide. So here are some resources that we’ve included for yourself. You can scan the QR code. You get a digital handbook that way. It has a lot of our key takeaways from this session. And as well as some of the presentation slides. And team contact details. So if you’d like to connect with us through LinkedIn and the like, we have all of that for you. Once again, thank you so much for your time and attention. And it was a pleasure. Thank you. Thank you, everyone.

T

Tayma Abdalhadi

Speech speed

149 words per minute

Speech length

1693 words

Speech time

677 seconds

Importance of interdisciplinary approach

Explanation

Tayma emphasizes the need for an interdisciplinary approach in AI education. She argues that discussions about AI should not be limited to technical experts but should include diverse perspectives to ensure a more comprehensive understanding.

Evidence

Example of the dynamic relationship between AI users and technology, shaping each other’s understanding and development.

Major Discussion Point

Teaching AI Fairness and Governance

Agreed with

Melissa El Feghali

Agreed on

Importance of interdisciplinary approach in AI education

Cultural adaptability of teaching methods

Explanation

Tayma stresses the importance of making AI education methods culturally adaptable and accessible globally. She emphasizes the need for grassroots efforts and effective feedback loops to ensure that educational approaches are relevant to diverse contexts.

Evidence

Example of classrooms with limited internet access and the need for tailored educational approaches.

Major Discussion Point

Teaching AI Fairness and Governance

Agreed with

Melissa El Feghali

Agreed on

Need for cultural adaptability in AI education

Teaching critical thinking with AI tools

Explanation

Tayma advocates for teaching children to approach AI tools critically. She emphasizes the importance of comparing information from multiple sources and understanding that AI outputs are not always the full truth.

Evidence

Suggestion to use multiple AI models and human sources for comparison and verification.

Major Discussion Point

AI Tools in Education

Agreed with

Ayaz Karimov

Agreed on

Importance of critical thinking in AI education

Adapting educational goals for AI era

Explanation

Tayma argues that educators need to adapt their teaching goals in the AI era. She suggests focusing on developing critical thinking skills rather than just memorization or essay writing.

Evidence

Suggestion to use AI tools to teach children how to think critically and creatively, rather than just for completing assignments.

Major Discussion Point

AI Tools in Education

Importance of feedback loops

Explanation

Tayma emphasizes the importance of effective feedback loops in AI education. She argues that users should be motivated to provide feedback and see the impact of their input.

Evidence

Example of providing feedback on AI-generated videos and the importance of seeing the impact of that feedback.

Major Discussion Point

Measuring Effectiveness of Interactive Tools

A

Ayaz Karimov

Speech speed

193 words per minute

Speech length

2783 words

Speech time

864 seconds

Using serious games for education

Explanation

Ayaz discusses the use of serious games as an effective tool for teaching complex subjects like AI. He explains that serious games have a specific learning goal beyond entertainment and can simulate real-life scenarios.

Evidence

Example of a cybersecurity game called ‘Duck Code’ used to teach cyber literacy.

Major Discussion Point

Teaching AI Fairness and Governance

Screen time management

Explanation

Ayaz addresses the issue of managing screen time when using digital educational tools. He discusses the concept of ‘cyber disease’ and the importance of limiting screen time, especially for children.

Evidence

Research suggesting a maximum of 40 minutes of educational game time per day for children aged 8-12.

Major Discussion Point

Challenges in AI Education

Testing different game types

Explanation

Ayaz recommends testing different types of games to determine which is most effective for teaching. He suggests trying various formats like board games, simulation games, and strategy games to see which engages students best.

Evidence

Personal experience of testing different game types with a biology teacher.

Major Discussion Point

Measuring Effectiveness of Interactive Tools

Need for AI literacy

Explanation

Ayaz emphasizes the importance of AI literacy in education. He argues that understanding AI, including concepts like hallucination in language models, is crucial for effective and safe use of AI tools.

Evidence

Examples of government initiatives in the Netherlands and United Arab Emirates focusing on AI literacy and prompt engineering.

Major Discussion Point

AI Tools in Education

Agreed with

Tayma Abdalhadi

Agreed on

Importance of critical thinking in AI education

M

Melissa El Feghali

Speech speed

159 words per minute

Speech length

1287 words

Speech time

485 seconds

Project-based learning in non-formal settings

Explanation

Melissa advocates for project-based learning in non-formal educational settings like scout groups and youth clubs. She argues that these environments provide flexibility for experimentation and foster collaborative learning.

Evidence

Examples of how project-based learning can be adapted to local contexts and AI fairness issues.

Major Discussion Point

Teaching AI Fairness and Governance

Agreed with

Tayma Abdalhadi

Agreed on

Importance of interdisciplinary approach in AI education

Lack of access to resources and expertise

Explanation

Melissa identifies lack of access to resources and expertise as a key challenge in implementing project-based learning for AI governance. She emphasizes the need for investment in infrastructure and partnerships to address this issue.

Evidence

Suggestions for policymakers to invest in digital infrastructure, create partnerships, and support flexible learning frameworks.

Major Discussion Point

Challenges in AI Education

Contextualization of learning materials

Explanation

Melissa discusses the importance of contextualizing learning materials for different communities. She argues that while core concepts can remain the same, examples and scenarios should be adapted to local contexts.

Evidence

Example of adapting a peace education program for different countries by changing specific examples while keeping the core concept intact.

Major Discussion Point

Challenges in AI Education

Agreed with

Tayma Abdalhadi

Agreed on

Need for cultural adaptability in AI education

Creating flexible learning frameworks

Explanation

Melissa emphasizes the importance of creating flexible learning frameworks. She argues that flexibility allows for experimentation, failure, and improvement in project-based learning approaches.

Major Discussion Point

Measuring Effectiveness of Interactive Tools

Agreements

Agreement Points

Importance of interdisciplinary approach in AI education

Tayma Abdalhadi

Melissa El Feghali

Importance of interdisciplinary approach

Project-based learning in non-formal settings

Both speakers emphasize the need for diverse perspectives and flexible learning environments in AI education, promoting a more comprehensive understanding of AI concepts.

Need for cultural adaptability in AI education

Tayma Abdalhadi

Melissa El Feghali

Cultural adaptability of teaching methods

Contextualization of learning materials

Both speakers stress the importance of adapting AI education methods to different cultural contexts and local realities to ensure relevance and effectiveness.

Importance of critical thinking in AI education

Tayma Abdalhadi

Ayaz Karimov

Teaching critical thinking with AI tools

Need for AI literacy

Both speakers emphasize the need to develop critical thinking skills when using AI tools and the importance of AI literacy in education.

Similar Viewpoints

All speakers advocate for flexible and adaptive approaches in AI education, emphasizing the need for continuous feedback, experimentation, and adjustment of teaching methods.

Tayma Abdalhadi

Ayaz Karimov

Melissa El Feghali

Importance of feedback loops

Testing different game types

Creating flexible learning frameworks

Unexpected Consensus

Non-digital approaches to AI education

Ayaz Karimov

Melissa El Feghali

Screen time management

Project-based learning in non-formal settings

Despite discussing digital tools, both speakers unexpectedly agree on the value of non-digital or limited-digital approaches in AI education, addressing concerns about screen time and emphasizing the importance of real-world, hands-on learning experiences.

Overall Assessment

Summary

The speakers generally agree on the need for interdisciplinary, culturally adaptive, and critical thinking-focused approaches in AI education. They also emphasize the importance of flexibility, feedback, and real-world application in teaching methods.

Consensus level

There is a high level of consensus among the speakers, particularly on the need for innovative and adaptive teaching methods in AI education. This consensus suggests a strong foundation for developing comprehensive and effective AI education strategies that can be applied across diverse contexts and learning environments.

Differences

Different Viewpoints

Approach to screen time management

Ayaz Karimov

Melissa El Feghali

Ayaz addresses the issue of managing screen time when using digital educational tools. He discusses the concept of ‘cyber disease’ and the importance of limiting screen time, especially for children.

Melissa El Feghali: I can add to that on what Aya said. It really depends on the game that you’re aiming for. A lot of the games that can be created don’t even need to use screen or have screen time.

While Ayaz emphasizes the importance of limiting screen time, Melissa suggests that many educational games can be designed without requiring screen time at all.

Unexpected Differences

Overall Assessment

summary

The main areas of disagreement were minor and primarily focused on different approaches to implementing AI education and managing screen time.

difference_level

The level of disagreement among the speakers was relatively low. Most speakers presented complementary viewpoints that enhanced the overall discussion on AI education and fairness. This low level of disagreement suggests a general consensus on the importance of AI literacy and the need for innovative, inclusive approaches to AI education.

Partial Agreements

Partial Agreements

Both speakers agree on the importance of AI literacy, but they differ in their approach. Tayma focuses on critical thinking and comparing multiple sources, while Ayaz emphasizes understanding technical concepts like hallucination in language models.

Tayma Abdalhadi

Ayaz Karimov

Tayma advocates for teaching children to approach AI tools critically. She emphasizes the importance of comparing information from multiple sources and understanding that AI outputs are not always the full truth.

Ayaz emphasizes the importance of AI literacy in education. He argues that understanding AI, including concepts like hallucination in language models, is crucial for effective and safe use of AI tools.

Similar Viewpoints

All speakers advocate for flexible and adaptive approaches in AI education, emphasizing the need for continuous feedback, experimentation, and adjustment of teaching methods.

Tayma Abdalhadi

Ayaz Karimov

Melissa El Feghali

Importance of feedback loops

Testing different game types

Creating flexible learning frameworks

Takeaways

Key Takeaways

Serious games and project-based learning are effective tools for teaching complex AI concepts like fairness and governance

AI education needs to be interdisciplinary and inclusive, involving both technical and non-technical perspectives

Contextualizing and adapting learning materials to local realities is crucial for effective AI education globally

There’s a need for increased AI literacy among both youth and adults to use AI tools responsibly

Feedback loops and grassroots efforts are important for making AI education culturally adaptable and accessible

Resolutions and Action Items

Implement project-based learning and serious games in non-formal educational settings to teach AI fairness

Focus on developing AI literacy programs for various age groups

Create partnerships between governments, private sector, and civil society to improve access to digital tools and expertise

Unresolved Issues

How to effectively measure the impact of interactive tools like serious games in teaching AI concepts

Balancing screen time and digital engagement with other forms of learning, especially for younger students

How to fully address the challenges of AI tools like ChatGPT being used for cheating in academic settings

Suggested Compromises

Using hybrid approaches that combine digital and non-digital elements in AI education games and activities

Adapting existing educational games and materials to local contexts while maintaining core concepts

Reframing educational goals to focus on critical thinking and creativity rather than just content production when using AI tools

Thought Provoking Comments

What I mean that it’s a dynamic relationship between the users of this technology and the technology itself. When we use the AI model, we shape it by the information we give it, but it also shapes our understanding of the topics that we give it and ask it about.

speaker

Tayma Abdalhadi

reason

This comment introduces the important concept of the bidirectional influence between AI and its users, highlighting the complexity of AI’s impact.

impact

It set the tone for discussing AI fairness as a dynamic, evolving issue rather than a static technical problem. This framing influenced subsequent discussions on education and policy approaches.

The main difference between the game and the serious game is that actually when you play the game, you just play for the sake of, like, getting entertained, just to achieve something. But in the case of the serious games, you totally have totally different ultimate goal.

speaker

Ayaz Karimov

reason

This comment clearly defines serious games and distinguishes them from regular games, providing a foundation for understanding their educational potential.

impact

It led to a deeper exploration of how games can be used as educational tools, particularly for complex topics like AI fairness.

Project-based learning doesn’t just stop when the activity ends. It creates a kind of ripple effect that goes on.

speaker

Melissa El Feghali

reason

This insight highlights the long-term impact and scalability of project-based learning approaches.

impact

It shifted the discussion towards considering the broader, long-term effects of educational methods beyond just immediate learning outcomes.

We need to go and see what actually motivates people to make an impact and to deliver their voices. And most importantly, how can we make sure that those voices that have been delivered are actually getting the counterfeedback they deserve?

speaker

Tayma Abdalhadi

reason

This comment emphasizes the importance of user feedback and engagement in shaping AI systems, highlighting a often overlooked aspect of AI development.

impact

It broadened the discussion from just education to include the importance of ongoing user input in AI development and governance.

If the goal is just to memorize or just to achieve an essay, then you’re failing as an educator to actually adapt with the technology. And the important thing is that you’re trying to teach your child how to have critical thinking using the LLMs.

speaker

Tayma Abdalhadi

reason

This comment challenges traditional educational goals and proposes a shift towards teaching critical thinking in the context of AI tools.

impact

It sparked a discussion on how educational approaches need to evolve in response to AI technologies, moving beyond concerns about cheating to considering how to leverage AI for better learning outcomes.

Overall Assessment

These key comments shaped the discussion by broadening its scope from simply teaching AI fairness to considering the complex, dynamic relationship between AI and society. They highlighted the need for innovative, engaging educational approaches like serious games and project-based learning, while also emphasizing the importance of ongoing feedback loops and critical thinking in both AI development and education. The discussion evolved from a focus on specific teaching methods to a more holistic consideration of how to prepare individuals to engage with and shape AI technologies in responsible and impactful ways.

Follow-up Questions

How can we create effective feedback loops between AI users and developers to improve AI fairness?

speaker

Tayma Abdalhadi

explanation

This is important to ensure AI models are continuously improved based on real-world usage and to address potential biases or issues that emerge.

What are the best practices for designing serious games that effectively teach AI fairness concepts?

speaker

Ayaz Karimov

explanation

Understanding optimal game design techniques is crucial for creating engaging and educational experiences around complex AI topics.

How can we measure the long-term impact of project-based learning initiatives on AI governance understanding?

speaker

Melissa El Feghali

explanation

Assessing the lasting effects of these educational approaches is key to refining and scaling successful programs.

What strategies can be employed to manage screen time effectively while using digital tools for AI education?

speaker

Audience member

explanation

Balancing the benefits of digital learning tools with concerns about excessive screen time is important for healthy implementation of AI education programs.

How can AI literacy programs be developed and implemented for different age groups and contexts?

speaker

Ayaz Karimov

explanation

Developing targeted AI literacy initiatives is crucial for ensuring responsible and informed use of AI technologies across society.

What are effective ways to teach children how to critically evaluate AI-generated content?

speaker

Tayma Abdalhadi

explanation

Equipping young learners with skills to assess AI outputs is essential in an era of increasing AI-generated information.

Disclaimer: This is not an official record of the session. The DiploAI system automatically generates these resources from the audiovisual recording. Resources are presented in their original format, as provided by the AI (e.g. including any spelling mistakes). The accuracy of these resources cannot be guaranteed.