Webinar Teaching & Assessment in The Age of AI

thank you for joining us today for the own Temple webinar teaching and learning in the age of AI we're going to give a minute or so to let people start joining us I can see the numbers going up right now I'm Isis from The Faculty of languages and Linguistics University Malaya and I'll be your moderator for today's special [Music] we're now going live so just give us a minute or so to wait to set things up while our participants are joining us we're now up to 213 participants and Counting [Music] [Music] [Music] [Music] foreign [Music] thank you [Music] foreign [Music] hi I'm sorry for the slight delay we're trying to work some things out with our live streaming platform um it might have been there has been a theme to a YouTube live link you can see that there is that it's printed the the Facebook link is the same thing as before but there's been a bit of it slightly to apologize for this unforeseen [Music] if any of you know of your network who is planning to attend via YouTube I hope you can help share their most the latest which you think which is in the zoom plan [Music] thank you [Music] thanks for letting us know what Dr yomi [Music] foreign [Music] here [Music] um again I do apologize for this paint in the YouTube live link um if you could help share this out to whoever that you've invited I would really appreciate that [Music] we will be starting a couple of minutes late okay so we're ready to roll okay so uh good afternoon to everyone and thank you for joining us at the own couple webinar teaching and learning in the age of AI I'm Isaac side from The Faculty of languages and Linguistics University Malaya and I'll be your moderator for today's session we have a jam-packed program lined up for you today with three speakers who will share their insights on how they use AI in teaching and assessment plus videos from two other academics some quick housekeeping notes we would love to hear all your comments and questions which you can send in via the Q a box in Zoom or if you're joining us live in Facebook and YouTube you can pop your questions and comments in the chat we'll have short q a breaks after each of the speakers and an open dialogue at the end of the webinar rest assured if we cannot address your comment or question during our webinar we will be compiling all the interactions after the webinar for further action so since we're starting a bit late let's get into it straight away I would like to invite associate professor Dr Levin demister director of the center of teaching excellence Singapore management University and chairman of our own Temple to deliver the opening remarks over to you Dr Lewin thank you Aisha hello everybody uh as Aisha said my name is Levin the maester I'm the chairman of the Aon thematic Network on technology enhanced personalized learning and it is my pleasure to welcome you to the webinar today uh in the Thematic Network we look for multiple ways to collaborate and one of the areas identified is the area of Faculty development a work stream led by mahidol University one of our co-lead universities and and so in this collaboration we we hope to share live webinars like these but also jointly produced videos and other materials that our universities can then use for faculty development the topic for today is teaching and assessment in the age of AI clearly I think there is a large opportunity to make learning more personalized with the help of artificial intelligence uh the long-touted promise of a computerized personal tutor May finally be coming into fruition and so with AI students can potentially receive answers to questions help with tasks and feedback on their work they may even get suggestions for learning activities and and for their careers what is also clear is that the role of the instructor will need to adapt significantly and I'm sure we will hear about that today in some ways it may be also become easier if instructors can get help from the AI tools and I think maybe we'll get some tips for that as well if so we can then have more time for personalized feedback to students so in different ways I would expect that this new technology will indeed make learning more personalized I look forward to learning more during the session today a big thanks to all who prepared for the session today and thanks in advance to those who will work on the resulting outputs as well special word of thanks for Dean Surin from University Malaya for volunteering last May to help organize the session and Dr helmy Norman from University Kevin San Malaysia for agreeing on the spot at that moment to be a speaker um it is uh yeah it's great to see that we have obviously we have now a wide program with multiple speakers before we start please allow me to hand the mic shortly to a collaborator from the University of the Philippines Dr Miguel rimalona our thematic Network also collaborates in the area of Education research and while we have you here we want to make sure you are aware of the research collaboration opportunities in the area of AI as well Miguel please floor is yours thank you very much Dr Levin let me just share my screen so there I think it's visible now so um good day again I am Miguel Francisco ramalana I am a faculty of the AI program from the University of the Philippines dilemma and I'm also the principal investigator uh for the AI research group of au mtpl so together with a group of around 12 researchers across various Au mtepl member universities across multiple disciplines involved in AI you would like to invite everyone here everyone who is participating um to our exploration on how AI can be used in personalized learning we have two main research questions that we are looking into and as shown here sorry I'm not sure my mouse is visible but if it is um it's uh essentially how do we leverage the tools that we have to enable and enhance personalized learning and the one we are currently looking into is that what is the perception in the acceptance in the use of these AI tools in the educational settings because once we have this information it's easier and it's um it's a we have more truth-based um Foundation as to build on how do we leverage these tools now um if you are interested in participating or being part of a research group kindly scan the QR code you can see on the screen for a Google form that should uh that would give us information more information about you and how we could contact you and we look forward to collaborating with each and every one of you so thank you for the time thank you Dr Miguel um for telling us about the interesting work that your research group is doing I hope many audience members sign up to participate in your study so now it's time to start our first presentation by associate professor Dr norisma binti Idris associate professor Dr norisma joined the faculty of computer science and information technology University of Malaya in 2001 where she's currently an associate professor with the artificial intelligence Department her research interest is natural language processing where the focus is on developing efficient algorithms to process texts and to make that information accessible to Computer Applications mainly for text normalization and sentiment analysis for the past five years she has published more than 20 articles on NLP and AI in various index journals she's going to speak today on the new era of artificial intelligence generative AI I would like to welcome Dr norisma to start her talk we will have five minutes after her talk for Q a over to you Dr norisma okay thank you so much the Aisha which are the same for today um assalamualaikum and very good afternoon um personally I would like to uh say thank you so much to the organizer for inviting me to to give a talk on the generative AI so uh since the period given the time given is very short so very limited so I'll share my screen okay uh so first of all before I proceed with the generative EI I would like to briefly uh share with you about the the Journey of of AI where um as we know that AI has started long time ago so uh it was started in 1950s where at that time the um the the aim of AI is to try to create intelligent machines try to mimic or replicate the human's intelligence so um at that time um since uh we have a very limited data and also because of the uh the limitation of the devices so most of the time we use um what we call as a rule base or pattern basic pattern recognition so um I was in the journey I started a venture into AI in 1990s so when I first did my research we normally use a rule base and we have to collaborate with human expert in order to get data and knowledge to acquire the knowledge from the human experts and so on so when it comes uh in 2000s it was the time where machine learning came in where um at that time uh there there are large volumes of data because of the growth of the internet so as we can see that from from that we have a lot of data coming in from the social media from the entertainment from for example like Netflix YouTube and so on from the forums like Reddit so we uh we have a lot of data and uh this massive data can be used to uh from that we can analyze the data and try to um uh get the pattern from those data so there was the time where machine learning came in so we have uh several uh models for the machine learning and then after that they try to improve the machine learning and uh in 2010 uh 2010s uh there was a time where uh we had a huge data set and then we have improved our devices and so on then we have a deep learning where the concept is the same as a machine learning it's just that the Leverage The the use of the neural network where they have several early years in order to process the data to do the training of the data and make decisions so um just a little bit on the uh the different domains of uh AI because uh AI is very broad so as we know that the aim of this AI is to imitate the humans intelligence so we have uh several abilities uh the ability to see to hear to comprehend and to think uh to do a decision to um learn and so on so uh that that is why in AI we have a different or a few domains so for example we have computer vision which is focusing on the image processing we also have speech processing which focus on the uh to recognize or speech recognition and we also have a Robotics and uh also a natural language precise thing so for NLP it's more on the processing of the uh tax uh whereby almost all the unstructured texts for example like all the social media uh we have like for example um the the forums so all this data has to be transformed into structured data in order for this data set to be a processed to be trained and to be used for decision making uh prediction and so on so as we know that AI has been with us for so many years whether it is generative or not so uh we have for example like Google translate um I'm sure that we use Google Translate and our students use Google Translate a lot and then we also have for example like um from aesia whereby they have their own chatbot so previously they had uh Ava so now they they introduced a new chatbot which is called uh bow then we also have like um facial recognition which are introduced by Facebook long time ago but they have suspended the features due to some uh security of the data security and so on and we also have like a Spotify Netflix where we have uh the features um what we call as a recommender or recommendation where we normally receive a recommendations of from Spotify AI on Netflix of the songs that our preference songs the genre of the songs and so on and I think that um when uh nowadays when we go to any uh probably the mama restaurants or even like uh the Johnny Steamboat uh we uh we will meet this uh this uh special robot which to assist uh the uh role of the waiters at the restaurants so all of this has been with us for uh so many years and uh we are used to all of the the Assistance or the the rules of this AI in our life so a little bit on the Milestone uh in the journey to generative AI so just now when I talk about the uh the Journey of AI we started with the basic uh pattern recognition with rule-based method and so on and then we move on to the machine learning where we have uh we analyze the data the focus is more on the analysis and to do prediction so we have a bunch of data we analyze the data we try to find the patterns from the data and then we uh sometimes we try to generate uh the insights from the data and uh we do make predictions or classifications from the data and then um uh after that we we also have deep learning uh few deep learning models like RN and CNN where at first the focus the main focus of these models are on the vision and speech where they are used to classify and detect the objects for image processing and for the speech they are used for the voice recognition to respond in natural ways for example like the series used in apple and so on and then uh after that um they will they're trying to improve uh the uh features or the ability of these uh these models um we using the same concept of machine learning uh they try to fine-tune some of the uh the algorithms so um there is uh there is a time when in during 2020s we have the language Mastery to try to uh um replicate our the intelligence of the our humans uh communication the language Mastery where a human can produce um sentences human can produce um designs or or data or images and so on so this is what we call a degenerative EI where the focus at first is more on the language based model language based AI models uh so that is why most of the time the applications are focusing on the which I bought like church activity and so on but and and then uh um after that they are trying to produce uh something for the images something for the the video voice and so on so what is actually a generative AI so uh objective AI objective models from the genitive EI the aim is to make the computers have the ability to create new forms of data so this is something that is different from the the uh the tasks of white machine learning and deep learning has done previously where um in machine learning and deep learning it is more on doing the prediction by using the the data that we have from the data set that we have we look into the pattern and we try to do the decision or the try to do the prediction but in this case quadratif AI it is more advanced when they are trying to make the computers have the ability to create new forms of data or create new contents so for this strategic AI it starts with prom so we have a problem that could be in the form of any text it could be in any it could be in image a text or it could be like uh if it is related to songs it could be in musicals notes um and a design and so on um and for destructive AI the most popular models that are normally use um the the first one if you are focusing on the image generation uh to do the image editing or maybe a deep fade creation which is uh what has been done and what have been uh some of the our celebrities probably the in Western the celebrities they have been the the victims for this creation data augmentation so uh for this um they are more focusing on the images so uh the the common models that I use against or genitive adversarial networks and there is another objective AI models that is also commonly used uh which is called Transformers initially it was used for uh text summarization chatbots recommendation in engines language translation or even in sentiment analysis um but uh now it is also being used for uh speech recognition as well as the image detection so for use cases for directive AI so as you can see that um these models can be applied in various use cases to generate virtually any kind of content so for example it can be used to uh in checkbox for the customer service and technical support it can deploy it can be deployed for the debates for mimicking the people or even specific individuals like the celebrities or even our um our business and so on it could be used for writing email responses uh to write resumes and project papers and uh it could also be used for creating the photorealistic art for the uh the creativity art creativity and so on and also for writing music and specific style or tone so one of the um the most common uh popular tool for attractive AI is what we are now discussing around the world which is a child GPT although there is another tool which is called but but I think uh for uh this talk I'll be talking about the check GT so it's it's one of the generative AI tools which create new content such as chat and also the conversation based on the problems that given by the users and uh for chat GPT it was built everybody knows that it was built by openai using GPT 3.5 which is a generative pre-trained Transformer um for two activity since it is uh more focusing on the written text so it utilize the large language model l m which is uh comes from the language model uh that I trained on a mass safe amount of data in order to predict next word to form a sentence n and M was previously from the language model where it is using the probabilistic uh coming from the hidden Markov model where by using this um model we can um predict the next word so previously we used language model in NLP applications whereby there are a few models which we call as a unigram diagram trigram and so on so it is depending on uh the number of words to be uh predicted so for example a unigrams only like considering only one word and then a biogram is considered considering two words and so on so the most commonly used previously it was by Gram and uh um previously Google Google Translate also use and gram as one of the models before they use a neural nmt so for this LMN because of the massive uh data set so they fine-tune the language model and it became a large language model and the use of these large language model is uh in order to respond to questions and to compose various written contents such as the Articles the essays idioms emails codes and social media posts so that is why we can see that like chatgpt they are they are able to compose um for example an sc one one specific essay compared to last time when we use large language model we can just predict the the words whether the next word or another two words so it is normally used in for example like email where when we want to compose an email we want to when we type in an email they will suggest a few words in order to complete our our sentence and uh for activity it is also trained by uh reinforcement learning to human feedback uh which use the concept of reward reward models and they rank the best responses so reinforcement learning is actually coming from the machine learning model and uh the fine tune it by using a human feedback so that it can be trained through the human interaction in order to understand the intent of the question uh and also uh we want to provide the most natural sounding and helpful answers so that is why when we use such activity uh for the limit uh they will assume that the answers given is uh is uh um so-called similar to to human uh answers okay since these are all the Advanced Technologies that we have here there are some concerns even from the smart people all right for example like Stephen Hawking uh he saw this uh this thing would be happening uh um one day right although he's he he um admit that the EI is our the biggest success in in human history but we will have to uh be very careful we will have to think of uh uh we will have to think of the risk that we are going to face uh one day and even uh Elon Musk also um uh raised the same concerns about this all these Technologies so uh the concerns surrounding generative AIS by example I'm just taking the example of the charity uh we know that everybody is talking about activity um and we know that our students are some of our students are using the activity so that there are some uh the rights of this um all these relative EI tools this is only charity we have so many tools for directive AI not only for text but also for other things like images and so on so um the rise of this generative AI has is also filling some various concerns um for example like uh in terms of the distribution of harmful content or biasedness um this is also the concern from from our our site from the technical background from our site from the AI people as well because um this objective AI they are depending on the uh training data all right so it could be uh cases where we have biasedness in terms of when they train the data so the bias could be in terms of the languages in terms of the content in terms of the countries and so on because they are just using whatever data that they could get from the internet so um this is one of the biggest ethical concerns and um also um about the harmful content because we are not sure of the uh the resources that they use in order uh to compile compile all these resources and to produce the the answers to our prom or to our questions so it will give a concerns on in terms of the accuracy of the uh answers or the accuracy of the information given by by these two we also concerned in terms of the data provenance all right the the [Music] which from where these resources uh came from uh when the time the timeline whether it is still uh valid or Obsolete and uh and and so on so um just just a tip on this uh where we although we use structivity but we have to utilize a variety of reliable uh sources when we're searching for the important topics so we can't just rely on this tool for accurate information if we want to have an accurate information it should not be a one or a sole uh single to for us to do our research and also the most biggest concern especially uh among our uh the Educators among the researchers is on the plagiarism so we know that uh this is uh also uh I have been talking by in all almost all universities about how we are going to avoid how we are going to monitor some of the instincts that came up with all the guidelines and so on but uh what I can see is that uh uh is is very uh General all the guidelines now are very general um yes we have uh we can't uh block our students to use all these tools because um this is what uh we have now so we have to leverage all the the tools all the Technologies so uh what we can what we can do is we recommend uh all of us to trade this tool as a additional learning tool but not as a mentor or reference to avoid the honestly completing all the their academic work so uh uh in the conclusion what I can see from uh all these what I can summarize from all this is that um the new era of AI is here we we can't avoid that we can run from that so what we can do is we have to embrace the technology but do not become it so uh um just let AI assist us in our life but uh um just make sure that we do not let it take uh over or become your life so that is why like for example like the the the one when when I share with you about the robot used in the restaurant um uh I didn't see that uh these are the the robots to replace the the waiters but it's actually to assist them uh in doing the task and also it we have to uh um to know that AI ethics is crucial and important all right so all individuals uh they have started they have started seeing the AI in their life in the jobs uh so um what I can see is that EI ethics become a very critical part of EI literacy some of the countries they have introduced uh AI ethics at a very uh early levels for example like in schools so it is a time for us to look at it and also to discuss about it and probably to make sure that in our syllabus we can embed about AI ethics so I think that will be the conclusion of my thought thank you so much thank you Dr norisma for the enlightening talk um as a linguist a lot of this is news to me and so I think and I think a lot of our audience members also learned a lot from what you've said we don't have any questions in the Q a right now um but maybe uh I think one thing that people are always wondering about the last points that you mentioned about plagiarism and you said some universities have their guidelines do you know of any specific ones that maybe um you want to tell us a bit about because I think that's what a lot of my colleagues are wondering all right okay um uh locally um I I know that some some of the local Industries um uh one of our local users are very fast but but still uh when I look at the guideline uh it's it's still a very uh General so so I think we will have to really sit down among all the uh among the the educations from the education side from the uh technical side from the ai's perspective uh we will have to look at because uh for example like opening ID they also uh aware of these concerns uh previously they launched what we they call as um is a tax return classifier whereby we can use that tool to uh know whether this return tax is uh actually human generated or machine generated but unfortunately they uh shut down the the tool because because of um the accuracy is still low so they can't we can't rely on the tool yet so they have to shut down the The Tool uh that is by open AI so probably we will have to look at uh apart from uh giving awareness um maybe we'll have to look at uh the guideline and some of the countries they even have the uh act for all these kind of things so so probably people will have to invite uh people from the you know for the cyber crime and so on in order to help us um uh what what are the uh then we can do after this apart from guideline and we the the most simple thing is to uh try to embed these AI ethics in our syllabus so that is what we did for example like previously we are trying to um uh give awareness about AI so we introduce uh the course AI for everyone so that all the the uh any uh backgrounds they can learn about a little bit a little bit on EI what is it all about what are the tactics we use and so on thank you Dr norisma okay we have a couple of questions here I will read out one of them because they're kind of running a bit late um we can uh Dr marisma you can stop sharing um so true Tom says it seems that we don't know the capability of AI at the moment to you how far can we go with AI and education I'm not sure you can answer that within a couple of minutes but we can try okay so this one probably I will have uh uh because um for for some of us are also involved in researching AI in education what we call SEI in education um for this one uh I know it's a little bit uh um three key these questions because as you know the period from for example from machine learning to deep learning it's just about a couple of years and then from the deep learning to genitive AI it's just about a few years so we uh we are not sure of um what what is the the advanced technology after this generative AI it could be more more advanced than generative AI where it can go further and it will uh impact our AI or our education as well so I think uh probably those professors here from education probably can can take the rule as well uh in order to help us in terms of the domain because uh we are just providing the technology but um but uh they don't mean uh the the impact is to The Domain which is on the education teaching teaching and assessment thank you Dr marisma okay I think I'm Gonna Leave the other question until later so that we can move on to the next part and you you mentioned law and that's a good uh transition for our next part of our program which is a video uh we're just getting ready to share it and it's a video from Jason schwa dean of the faculty of law University Malaya and he's going to share a bit about how he uses chat gbt in teaching law very good day um to everyone my name is Professor Jason Chua and I'm the dean of the law faculty at University of Leia I'm very grateful to Professor Surin for the invitation to share I suppose uh our own practice in using chat GPT or other AI systems in the delivery of our legal education curriculum so um I know that my time is short so I won't spend too much time rambling on with our introductory or Preparatory remarks and I will now share my uh slides with you ah right so um I have entitled my little um brief session chat GPT as a learning Tool uh for law uh please excuse the uh typo on this slide right so um of course as any self-respecting academic who uses chat GPT to teach I had asked the same question of chat GPT and this is the answer that I had received from chat GPT it tells me that I can use check GPT as a learning tool in providing guidance on how students can answer questions I I can use check GPT to clarify Concepts I can use it to assist with legal research for my students students can use check GPT to help with their grating abilities and competencies I can also use chat GPT to help my students appreciate the finite use of language check GPT also comes up with ideas and tips and hints on how to study and how to use one's time very carefully check GPT can also serve as a discussion partner to bounce ideas of one another and of course check GPT according to chat GPT is also a counselor being able to provide some uh advice and guidance on mental health and well-being and of course last but not least check GPT could also be used to develop games and quizzes Which could help with the learning process I'll just pick on one or two of these to explain how we use it in legal education so I find that answering questions is a really good one because that is perhaps one of the more axiomatic or one of the more uh uh more intuitive uses of chat GPT for students they are face with a tutorial problem or when they are encountering a coursework question bang they go on to check GPT type in the question and they hope to receive the answers generated by chat GPT now I'd like to use that that motion that that process to help students look at the quality of the answers uh uh generated by check GPT so what I would say to them is uh generate an answer based on this question now I want you to evaluate the question and Mark it as if you were the examiner and after marking it after assessing it I want you to provide um a very short one-page uh summary of your so-called assessment criteria so what were you looking for when you were marking this piece of work that Mr or Miss chat GPT had produced so another good way of using chat GPT other than um as providing a vehicle to for students to undertake some degree of applying the assessment criteria is also to use check GPT for research and this takes me to the third point which is about assisting with research now one of the challenges I find especially postgraduate students is that they actually don't know how to search properly for Solutions or resources relevant to their research problem or their academic problem so what I like to say to my law students is that look you have got a uh a long problem type question a scenario based question for example it may be based along some very broad Banks whereby a had taken a piece of cake on the table without paying Forex and and now B is unhappy with that and B has a reason or to uh claim that this is theft right so those are the bad facts now if let's say the students were to find the answer to that scenario problem through a chat GPT I would ask him what would you search for well many would say oh I'll just type in the uh problems set I would say to try that and if they do try that they would realize that the answer produced is really quite deficient so I like to ask them to practice asking the right question and I think that's really important long gone are the days when we would go to the library and look for the relevant books through some kind of uh Library catalog now the expectation is for goods students when trying to research uh to apply the correct Search keywords and not only keywords when it comes to artificial intelligence driven packages but also I think key phrases he questions so I like for them to therefore demonstrate that flexibility that fluidity of language when undertaking their research through check GPT so these are two ways I like to use check GPT for legal education now of course I could spend more time talking about some of the other uh uses of chat GPT as a learning tool but my time is up and I shall therefore stop sharing my slide and bid you all a very good afternoon and all the very best thanks very much we would like to thank Professor Jason schwa for sharing the video if you have questions for him he's not here but you can always pop them in chat and I'll definitely pass them on to him okay so it's time for a second speaker associate professor T.S Dr Muhammad helmi Norman Dr Helen is an associate professor of digital and futuristic education at the faculty of Education University kabang San Malaysia or national university of Malaysia he currently serves as the deputy director for instructional Technologies at ukm and is an apple distinguished educator he consults on educational technology transformation at higher education level for the ministry of higher education and at school level with the Ministry of Education and at Apple southeast Asia I'd like to welcome Dr helmi to present his talk called unleashing The Power of generative AI in teaching and learning for higher education over to you Dr helmy thank you all right okay so let me share my slides and I'll kick off right okay so thank you very much everybody uh associate professor uh leaving and um Professor Surin and uh and everybody at Allen Temple I think excellent job uh in you know organizing this and also uh the host uh University Malaya Singapore management University uh punai and also my fellow panelists awesome it's been it's an honor to be with you guys and um yeah so let's kick it off unleashing the power of generative AI teaching and learning for higher education all right so I'm helmy and uh national university of Malaysia and so what and why AI all right so this is the framework that I drew based on Professor Schwab's first Industrial Revolution uh definition so he states that you know from the world according Forum so he states that physical it's the 4r is the transaction between physical digital and biological worlds So Physical you know Technologies uh including autonomous cars uh drones robots and digital world metaverse vrar and then in the biological uh World synthetic biology wearables brain age you name it right a lot of advancements of technology and these worlds intersect to become fire but where's the AI the AI is actually the brain of all these Technologies so it's really the cognification of things so um that's what he said I'm quoting him excellent Peace by him you should read his book uh and so today we'll be talking about digital world or fourir and specifically the popular or trending viral generative AI now we're talking about that so here is the transformative um diagram of AI and then specifically I tap into generative Ai and it stated that these teams will be related to genitive AI health and Healthcare future of media and sports future of computing education skills and learning now we're going to talk about that today future work economic progress arts and culture and also civic participation so again this book UNESCO released this in 2021 before this charge gbt and everything but this excellent piece of book I think you should have a look at this guideline if you're looking to develop some policies on AI excellent uh talking about you know emerging practices benefit risk assessment and then moving towards sdg4 for that purpose yeah and so again this one is I think a lovely piece by um Spanish researchers here on the most popular genitive AI now trending right so so text the image you know Dolly stable diffusion Muse text to 3D image to text text to video text to audio text to text text to code text to science other models a lot of a lot of things that are happening it's not just GPT but gender ni uh in general right and the recently you know release chai gbt uh UNESCO uh and AI in higher education but I would love to you know highlight this section this is a great section where it talks about the roles uh you know you know that that you can take as an educator in higher ed so how could you use this possibility engine you know Socratic opponent for instance AI acts as an opponent to develop an argument collaboration codes uh guidance side personal tutor co-designer Exploratorium study buddy motivator and dynamic assessor so moving to that but you know generative AI is is actually just a tool if you don't know how to use it it's not going to be effective it's not going to be meaningful so I think it's important to craft prompts that are good and effective I would like to decide uh Barrett I think a lovely piece of a high quality craft prompt um he uses the create form framework so it's C4 Clarity so defining the task or intent uh putting that in your prompt and then relevant info uh you know providing detailed more the more detailed prompted the better they are can generate for you so examples if you put some examples there they it will generate a more good high quality uh output and avoid ambiguity so unnecessary details don't put it there and then Tinker so you you can generate a lot you know like generates like nah I want to generate again generate again and finally evaluate um as as needed and in and improving the quality you know Tinker again evaluate Tinker again valid so I I always say that you know like AI is actually like a research assistant it's like my virtual research assistant you have human ones and you give like uh what do you say instructions to them right so this is the same thing even though it's a it's a bot it's it's AI you have to give good prompts you have to give good instructions for it to go and give you good output so um uh have a look at this uh framework awesome framework so give and give some examples we go before we go to some awesome tools uh so these are some examples of generative Ai and metaverse I'll show you uh first of this and this is this is a course at um with with uh with the faculty of the education at my uh at my faculty so what we're doing there is this is the course of doctrine of Education um this these are so students from Malaysia and and China overtaking this course and so um I told them that uh refer to the UNESCO guide on jgbt look at metaverse and how do you develop this for your uh teachings at the University so they were doing for cabin service so let's have a look we use AI to serve a conversation situation passenger yeah just attendance there is a German passenger on the plane who does not speak English and I need to say hello to him what should I say couldn't talk it means good day or hello in German as a flight attendance there was a two-man passenger on the plane who did not speak English and needed to be asked where his lead was which means where's your seat as a flight attendant there is the two-man passenger on the plane who does not speak English and needs to be guided to her seat what should I say the physical body all right so the video is a long so um anyway uh I'm just showing you how you know you link the chat gbt meta versus you know VR AR all this you know emerging Technologies into that and and actually uh we can use it for very powerful instances here you can see language learning happening and and you know uh uh in terms of communication I don't know how to speak German but um okay now I know a little bit of German okay and now looking to jgbt and AR game development so this is this is a recent project uh with Malaysian Development Corporation so uh we teamed up with mdec and also this one was a a course um well it's actually a program for school students and this is the game develop and look at what I did with chapter GPT ndr games and this was only developed with an iPad let's have a look so I'm tapping uh so linking to AR games using chat GPT and the theme was UNESCO sdg and achieving 15 so I got the game design there foreign [Music] cool reform it's too named we have a relief conscious nature according the same thing foreign [Music] but um what I wanted to show there is actually there's a lot of things that we can do I had no idea oh my God what I'm going to do and I started to use chatgpt and in brainstorm of course the you know the design is uh co-created by me and Chaturbate but again um the you know where to put that and you know actually designing the game was me so this this this is the shift that we want to show you um on you know how to use chat GPT in a positive way in an ethical way uh as the the the the first um speaker was talking about yeah all right so now very exciting a guide for professors tools for teaching so I've uh matched up a uh what do you call it a a padlet okay for you so you can scan this um quickly and I'll paste it later uh in chat so just go to https uh bitly so bit dot bit.ly slash on Temple 2023 and you'll get some awesome stuff there okay so I'm gonna show you all right so um this is it this is the uh guideline the padlet so I've put it on on the left there resources and you know I I said is coming out with their guidelines soon uh some some prompts there let's go to have a look at po.com not gonna discuss anymore but pro.com so it has the you know like mashing up of all these genitive AI awesome text to uh text to text and this is um text to image so um stable diffusion is free uh everything else I think you have to pay for it but um awesome um stuff on stable diffusion all right so let's have a look um on stable diffusion I've put a prompt here um to for the AI to front some image some generative AI so you know um this could be done for storytelling this could be done for engineering this could be done for medicine you know there's a lot of opportunities there's a lot of possibilities here but um during this so what I did is prompted a you know sci-fi City I wanted that that picture so it generated pretty pretty good for me yeah so have a look at stable diffusion again um it's on the website that uh I showed you the padlet all right so this is awesome as well this is uh blockade Labs um this is text to 360 images so what I did is prompted for I think a beach in uh tropical beach or beach at night at uh with corals I think so um and here you can see here um the digital painting you can you know change that and actually this is a Skybox you can uh plug this in in uh Skyblock environments uh for instance in in metaverses you can use it as a 3D World so you can you know overlay it so it looks like um you know you're in a space I think this is awesome um you have the long you know long long long you have to you have to have a 360 to put in AR VR um and and you know especially if you are put the goggles on you you feel immersive so um these are a good or interesting uh uh you know platforms for that okay so moving on so I do eight so this is awesome as well um text to lesson planner or you know like uh generating a course syllabus I was like oh this is so awesome and I put here um create a course on metaverse so I I prompted this in in this uh generative Ai and it magically placed the lesson for me so look at how we can do of course we are the experts the the I the AI again is isn't is a research assistant for me so um it gives me it gives me quickest assess and of course I'm not gonna you know like I can't copy this but it's called plagiarism right so we we copy or or we use this as a prompt for uh good brainstorming and then maybe some other uh you know AIS mash it up and becomes are so I think um it's the same thing if you look at you know um examples that are out there and then we you know sort of take five or three out of them and then you know like refine so it becomes harder so that's the thing that we want to teach uh teachers Educators uh students um to to not or use this ethically yeah so and listen let's have a look at listen so this is tools for research just now tools for teaching but this is awesome as well um I've tried this and and I try to put uh so it it generates a literature review systematic literature as well oh my God this is so awesome so I was uh prompting it uh mobile learning or mobile learning wherever you're from right uh mobile learning modeling in Malaysian schools from 2010 to 23 so in in a split second I was like oh my god oh this is so awesome and and you know and you could generate the systematic review like uh I mean uh magically and within a few seconds uh well again uh the AI is not perfect yet but I think it's going there it's moving there um excellent excellent stuff uh again uh elicit have a look at that uh for research too so uh before this I showed you a bit on teaching tools and now moving to research tools and then here is another awesome one this is called connected papers so it's text to article uh visualization I love this one so I put on uh prompted metaverse and then just click the connected papers and I wanted to have a look at this one um you know so all these connected papers um AI generated for me and I was like oh this is so awesome this is and and you know you could you know easily um create papers after the and and again if we're looking at a at a negative it will become negative but if we look at I generative AI from a positive side it will really help me and again ra at our virtual ra i disposal awesome awesome stuff by connected papers okay so I think um to recap um what's next for for generative AI um actually we are in Malaysia uh we we are working with the uh all the universities um in the Malaysian e-learning Council for a higher education which I'm a part of uh we actually came up with the framework for AI and and we're developing a guideline um for AI in a higher education uh okay hopefully you'll come out soon uh but yeah of course I think the key Point here is to look at governance of generative AI at institutional National Regional and international levels I think if our own uh Temple would have a go on genitive AI uh have a guideline I say awesome it will be like awesome stuff right I I know uh Asen is is coming out with 2024 we'll have a look at that and then we'll see the Caps there and and you know um fill fill it up right policy guideline development support system of course is important to understand that and stakeholder involves we have to of course uh you know make administrators and support staff understand about jtpd as well it's not just about Educators and also Learners so we have to have that as well and teaching and learning so of course moving uh to pedagogical design delivery of course before this it was like how do we you know design for online learning how do you sign for this and I was like how do we design for metaverse and I was like now is is the arrows like how do we design for generative AI so it's it's very exciting of course daunting as well a lot of uh opportunities and also threats uh um learning design and delivery of course assessment and evaluation uh involving Educators and Learners of course again this is a very important integrity and ethics uh looking at uh academic intimacy learning ethics research and Publications and of course uh finally going for uh training and professional development of generative AI looking at AI literacy intensive AI tools and of course um all the stakeholders involve uh so my final um this is the quote from chat GPT so I was like asking give me a quote for my talk today on AI in higher education this is told me I was like oh my God so awesome so as AI continue to weave its way into higher education do we find ourselves at the brink of transformative era where technology complements the brilliance of human Educators so embracing AI responsibly can unlock personalized learning experiences Foster critical thinking and Inspire creativity among students preparing them to face the challenges of tomorrow with confidence and in Ingenuity I was like oh my God so um this is chat GPT this is generative AI have a look and think how you can integrate in your teaching and learning so I'm help me signing off back to you Natasha thank you thanks so much Dr Hill me and also for sharing your paddler with so many uh interesting tools we can use uh it's hard not to get inspired by Your Enthusiasm but I'm gonna throw a spanner in the works here with a question from one of our attendees who says yes AI also can do wonders but how about the impact on evolutionary natural intelligence Primal skills and psychological well-being that seem to be diminishing among many we need more discussions from learning and developmental philosophy perspectives if we want to inculcate holistic human values through education while embracing technology otherwise we will descend to the Hunter and gatherer being what would you say to that well that's a long question but thank you for the comments I think yeah of excellent uh views definitely I think of course um with with this new you know um technology of course um we're I think it's good to I like to um you know uh always quote um Professor he's a psychologist Professor uh Bruce tuckman um so he has this four phase of of forming something it's the first phase is forming and then you know you you go into the face of storming so he's like oh my God how do I do Zoom how do I use chat GPT how do I use generative Ai and then the third is actually the face of norming where oh that's easy oh you use this oh you tap this oh you use my pageant anyway I'm saying like and then finally you go into the phase of Performing where you understand the ethical uh you know ethical implication this is what I can't do this what I can do so understanding in your Niche area in in your discipline I think of course that's important um and I don't think you know um uh the the the teacher the educator the professor will diminish the rule but the rules change of course before this again um Sage on stage and then now you know facilitator may be learning Inspire and then you know chat GPT coming in yeah yeah you can become the AI assistant you can become my research assistant so I think the role I think understanding our new role uh Advanced role I think is important in this era of AI yeah thank you Dr Hill me you mentioned research so I'm gonna like move on to a question we got in YouTube and it's from city north soleja harun and she asks is the chat is the citation list produced by chat GPT is it reliable to use in research okay I think again it's it's like um I always like to humanize the technology for instance I'm saying GPT is actually a um it's an RA research assistant of course um sometimes are a you know do mistakes they're human of course AI does mistakes as well so you are the expert don't don't let technology be the expert you are the expert you are the one who knows that it's true or not so you check it I think um that's the that's that's how our rules change because we are the top tier expert we are the professors so do not let AI profess you so you are the professor so I think that's important for us to think about that um verifying information across the rules change but I think um yeah important to to check the reference of course to gbt and other uh I think tools that I showed you you can use that as well but of course you have to double check on that uh to to check on it is it on its uh validity and reliability yeah thank you I have another question here uh with the fast development of AI is there a possibility that some work or jobs will become obsolete especially in the field of Education very very interesting of course Ah that's a hard question to answer I I think uh again um rules are changing and of course um jobs are changing um of course you have if if you remember the diagram I showed you for IR right um so autonomous cars so um is the grab drive or or maybe like um it's the taxi driver not gonna like work anymore no the change to a grab driver it's like Uber driver so I think in in education we change our role as well not as maybe not as a content provider anymore but a Content facilitator or or you know like a learning facility so maybe you can search this you have um you know this form chat GPT from AI is this is this true is this reliable I think uh of course the roles will diminish um some rules but of course it will open new rules you know new rules new you know uh who's going to become the AI programmer who's going to become the uh you know GPT analyst but there's a lot of rules uh coming up and I think um that's there that's our role as an educator to prepare students for the jobs of Tomorrow really think how we teach uh now in teaching and learning and and prepare them for for the uh volatile uh learning or volatile uh Workforce of tomorrow thank you Dr helmy okay so um there are a couple more questions but we're running out of time so we have to move on thank you so much for your uh interesting talk um to those who have asked for uh Dr yomi's presentation and reference list will update you on how they'll be shared uh via the chat uh but for now we're gonna move on to our second video schools our second video is on chat GPT assessment design Helper and it was created for us by Dr zahiruddin featured in Abu Hassan he's the head of training at the academic enhancement and leadership development center or adac at the University of Malaya we're going to share the video now foreign today we are going to look at several prompts that can help you use Tech GPT for assessment design don't worry if you did not manage to catch what those problems are I have listed them down in the video description I'm going to use an example from the course that I have taught this semester which is construction law advice you need to give check GPT as much context as possible in order for it to generate the questions that will match your needs as close as possible so use my problems as a guide or as a starting point for you to start designing or start thinking about what kind of assessment that would you like to give your students and also how Tech LGBT could help you to diversify the kinds of assessment tasks or questions that you give your students I'm going to start with the first prompt that I've already created for this video and this prompt will look at the continuous assessments okay so I'm going to copy the prompt that I have okay it's uh generating but if you look at the prompt here you are giving contacts in terms of the learning outcomes and also the cost synopsis okay so now you can see that the activity has come up with one two three four five six seven assessments that you can give your students okay so it ranges from role play scenario even list the learning outcomes that it covers okay a group presentation problem solving exercise openbook quiz legislative research okay so that's the examples that adjectively use you can also try to ask another question so um probably I would say that the any of the questions given for the proposed question it's not really matching my needs so I can ask the expected to come up with another one example like this okay so the quest The Prompt ask the scientifically to come up with five non-conventional out of the box group cost for the assessment for construction law okay and I excluded legal research because I don't want the chat GPT to suggest legal research for this okay so then they will come up with a new list again gives you the title of the assessment and what are the descriptions and what are the rules for the students so comes up with five so let's say for example I like the construction law debate tournament so but I want the check GPT to still help me with the design for this so I'm going to paste in another prompt this this time I'm going to point it directly to question number four okay so I'm going to change this such as the table format evaluation rubrics to option four because four is construction log debate tournament okay and I asking it to give the rubrics for this so now you can see that check DPT is creating rubrics for the assessment and I limited or I've told jhbt to be specific with the rubric okay in table format so it gives me a table format like this and then limits the Criterion for each of the rubric category to two and rating scale from one to four but I don't really like this the way that it generates this the table format so what I can do is actually just click on regenerate okay and then on the left hand side here you can see there's two out of two that means you can actually go back to the first one if you take that the second one is less interesting okay so again so this is the rubrics test given by gbt and then it also gives this description on what the assessment will be for the students okay so that's interesting isn't it so now let's look at how we can use jgbt to give us a a bigger kind of evaluation rubrics for assessment so again you want to be specific with activity to give it context so this is another request for rubrics okay so in this request for rubrics you are asking for two categories of rubric ready scale one to four okay but I would say that this is too short for two rubrics so I'm going to stop generating here and then I'm going to tell GPT to change it so by clicking on this so I'm going to change the Arabic category to five now okay and then we let the relative scale to remain one to four click save and submit it will generate a new rubrics okay and now it has four criterias Okay so simple is it so let's look at how we can generate exam questions for sorry okay so we can ask a general question like so okay so remember that we've already told jgbt very early on just now about the learning outcomes okay and using this chat chat activity knows that learning outcomes number one is elaborate roles and objective for construction law number three apply conception law in contracts Innovative problems so it will ask questions that covers both in the outcome together again it will give you the example questions for you to work on so it doesn't have to be that you have to use whatever we get if they gave you you use it to give you context you use it as a starting point so that then you can change the questions as how you would like it so let's look at question number five okay so how do we then create a marketing scheme or marking guide for our questions based on each activity so remember that check DPT actually is not really good at domain specific domain knowledge but we can still use it to inform us on how the likely answer will be okay so I will ask 10gbt to so I'm going to ask you to ask answer the question number five so now in a limited manner so we can look at how Tech gbt answers the question so you can use this to start off your marking guide for the exam question so lastly let's ask the GPT to help with help us with something that is I'm not really something that we are not really um normally do we know something that we know we don't normally do especially or at least on in my field which is open book question Okay so I'm going to paste this prompt okay so I'm asking check gbt now to create an open book question okay and this is on a specific learning objective so remember context is there so you give it context Again by putting in University student okay and also uh what you want the students to comment on and also create a make believe case okay so so what GPT gave me actually uh is an open book question but there's no make believe case let me ask you to regenerate the question again okay so now it gives me the case that that question should come up with so that's how you can use check GPT to help you prepare your exam practice or to prepare your assessment continuous assessment question hopefully that helps thank you very much okay I hope you enjoyed and or learned something from Dr zahid's video we will be sharing the link to the loom site where this video is hosted so you could uh interact with him there um it's now time for a third speaker I um like to invite Miss Sheila yusu Che us she graduated from the American culture and literature Department in 2003 and then pursued her master's degree in teacher education in and she has been a full-time instructor tired from somebody University in Turkey Sheila is interested in the implications of AI on higher education permaculture neuroscience and neurosystemics care I'm gonna hand over the floor to you now in Sheila thank you very much okay thank you very much connection it's very nice being here and so that I'd like to thank dear Esurance um okay uh today I'd like to talk about empowering polish businessmen uh by I'm missing the potential vocabul um I will start first of all by talking about the background but I won't really go into detail because doctor doctor nurit Mahari uh made a verbal instruction for it so you're familiar with it already and I'll talk about concerns related to chat recipes and then I will talk about how tattoo people can they bring about the tournamental chip in higher education systems okay I I first of all I also have a background uh I asked chat to PC to come up with an analogy okay chapter 250 came up with this analogy uh metaphor and uh in terms of using touch UTC in higher education and this is what it came up with so sorry is similar to astronomy knife because both of them provide a range of functions in their contacts design both of them awkward and all-in-one content for assessment uh what does that operand all in one courses and both of them combine experience methods and techniques to process diversity however we need to utilize them properly in order not to tell myself okay uh now I've asked for consider the concerns related to [Music] um generative AI the first question you can imagine is academic integrity because uh linked AR using specifically students qualities and strategorize especially those writing essays and you can ask students are very creative in using their AI they may add short GPC to write an essay without being detected without realizing so it is not really possible to discuss the other context this is one of the biggest problems of health agency uh we as prevention we use permission and now we have a certificate plugin but it can um like 100 I'm ensuring the Institutes of Child Support assessment is challenges okay for this reason academic interesting one of the biggest concerns regarding generation AI the second concern is related to buy and fairness and accuracy chapter C stands for generative pre-trained Transformer well it has been trained by somebody previously but we don't know who trains have to treat me okay for this reason we need to see right there uh so animals May inherit viruses well today for their so on and we don't know what kind of bounces they may have uh this thing is for potential discrimination and answers in evaluation and when we talk about accuracy experiments so while using pregnancy we need to pay attention to whether the information is buzzed the other concern is related to basic security analysis they are more mustaches this week how did benefit to memorize a concept training data leading to information research when you first can detectively it asks you that the information uh will be you trained as if Affairs it kind of has the information you because the important thing here is over time we ask a question to test it we can never engages this but we kind of train ourselves so this is important and authorized access could be accessed by militancy leaving subscribing to violations and lead to those personal information it is possible for example you can answer your students information to be able to do research for example or people to use personalized uh feedback you may answer your screen data in chat easy but uh you never know who's to access the information uh that's your answer so we need to also pay it into it bowling and supposed to been tested students are requiring low-box measures to safeguard information uh I will not go into blockchain um technology but it must put ideas to Bear blockchain uh measures in mind we use test KPC because it's not going your way the other concern is the latest particular discipline and AI systems reflecting human education is a successful job displacement he thinks that while if we have a machine that does not sleep that that is okay it's not possible for us to outperform it and improve uh but I'll talk more about uh generation but this reminds me of the blue guys Oxford Dictionary those number of English workers uh who discovered Machinery especially in cut and wooden Mills thinking that they believe for accepting their jobs well it does not this reminds me whenever I was in capacity PC I always think about the blue diets because some universities as you know bands have 50 and generated AI uh thinking that students is not really possible to provide valid and assessment integration sector BC uh so if you ask me uh we will just see if we don't have we will be like the lewd eyes trying to destroy and injury uh isn't it true that some of their job was it safe is true by the way but uh it is not going away you need to find ways to integrate practices in first curriculum but we also need to be very careful uh as well so we calculated destroy it we can't freely make its way we need to find ways and recreate that integrated uh bearing the concerns in life another concern is related to limited subject matter so when we talk about what this chapter and the nursing AI they are Advanced footage for objective and multiple practices okay you can ask just to grade your students multiple choice for you for example they work well however they have limited Effectiveness evaluation subjective or creative here is uh for example why are calculate processing skills how to order those things instead um methods to resort to okay here we need the humans here we need the teachers to tackle these sort of skills uh at the moment uh how can we I mean how to decision a transform hiring status is concerns and those concerns are really serious but few people can also provide us with some server lines okay and to be able to align ourselves with evolving technology to ensure reliability and validity we need to reduces to convenient more services students before testing people they personally did in this case we need to think about ways in which we can make assessments more reliable and okay I'll put we can use to thank people I'm most of the universities are already uh music we need to provide our students with real work allowing students to expire their knowledge ask yourself in Practical situations you need to promote critical thinking too because if our students only use Google trust gcp there is no way they can use to focusing skills some of my students because there are still used to using Google another generator for example while during the pandemic we were doing on the own room and one of the questions asked dear future with us I don't have the electricity at someone so there is no way I can connect to the lesson okay welcome to use your phone to connect to the session okay so the critical thing is still uh are really very important I recognizing a reliant investment methods we need to provide project presentation case studies stimulates role studies and we can get protective disease help when we incorporate these assessment methods in private curriculum as ongoing observation of students progress very important when you deal with profanity testing and thought churches an encouraging collaborative and interdisciplinary tasks we should need to be aware of different disciplines and they need to know what's going on in different uh disciplines this is very awesome this happens in this aspect of that reflection we should provide our students with continuously we need to get qualitative feedback promoting feedback we need to do research and respect ones for a fancy system to work out and uh let's show you some point which I think you're familiar with uh you know give your spirit a science project where students design and conduct experience for a real world issue for example we have a place called hope where they can make a three-dimensional object using 3D printing for example okay they really enjoy a group presentation on a social issue for example I know that University of Mala is very interested in you have a sustainable development goals you can give them a goal they can't work on it they can do a research using tested captured people so you can have a look at the uh issues on different lenses and they can work on their communication skills a business simulation where students like to do them as as counter negatives to solve strategic challenges uh you can also provide things for the portfolio of creative writing people artwork or design projects showing students parents and growth over time uh I would like to show you how to access the farthest Savannah University we had a very portfolio it is a graded component okay this is a Connect oh in class and then you come up with outside password these people can't compare students online with if they're in classroom so first of all we write contagious instructions for example they learn about plastic okay they receive that input inside for like weeks or two weeks and then learn the writing and specific conventions requires products for example when you get raffle we deal with service process if there's us about some social process they learn about how to write this process they arrive essays or record videos out of the giveness question can be a statements and then based on the outcome they have tutorials with their students discussing their progress and stuff in their sons and weaknesses and in this case for the first they can receive a yes or at partial if the outcome is that yes they don't need to read this path but if the outcome is a partial they may think through the path okay exactly they are asked to respectfulness Performance formatively and summative we see some uh like this uh basis on a diagram simply let's say a two three paths and respect weaknesses in terms of those tests while you're testing on the past please pay attention to being the easiest people hi I thought um we can make students as part of the learning and Grading process here uh we have a certain uh format for learn to Associates but this time for example students can't deny as structure their own learning portfolios we can give them the flexibility we can this will help somebody entertainers or for example they can offline uh the three sections they can decide on the order of content they can receive suggestions for for students their achievement and capture system also have reflection and subsystems it has constitutes with questions that include self-assessment practicing them identified as trans angry another thing that can that participation houses is creating a favor paragraph This is one of the biggest concerns when student cite essays we are scared that uh please look later right and it's a valid uh concern however I would like to offer an authorities here are FPS a social and political science teachers are trying this out at the moment that's going to be a lot uh we found first of all this could be about their essay or the part of community group we can give them to come okay we can under question that they will write you see and then they can research evidence and they have other people have approached using plasticity okay here's your work may come into play as well and then they can analyzing having analyze the questions how do we analyze different lenses uh based on decide what their arguments are going to be here they may use testic or they may use collaborative rather than yourself and then they can write their essays but the essay should be student owner how can insurance students May write this in Christ okay otherwise it's not really possible you know uh if it is almost but this is another this is like a suggestion to integrate entropy into life uh other speakers have mentioned this as well uh I asked practically speak to come up with this criteria to come up with those look rich okay you kind of say okay for this type of suggestion might be other people's professors have mentioned you need to be very specific when you contacted you see so I said can you please test 56 provider rubrics for essay writing uh I'm I've provided very specific guidelines and this is only one this is understanding the topic and we started okay so usually include testicity interface to see we can also we also need to update um our rubrics we need to consider the criteria um we also uh practices can also help us understandability and efficiency uh because teachers only lost 20 percent of what we've taken now we do a lot of math teaching work and FCC uh chat box that has large number of assessment simultaneously reducing the burden on substances can be this may not be done for all the assessments uh uh kind of um class box for example multiple a choice quizzes mechanical equipment uh passports can help us a lot in that sense and they can't go to library certain States and even as capital bullets areas to come to the courses and everything you can just not be helpful with that uh first of all to be able to do that teachers need to familiarize themselves with the AI now we need we don't need to fear them but we need to learn it until we can do that from them we need to define the learning objectives and understanding the specific skills that always that students need to demonstrate we need to think about the outcome uh if you would like to integrate AIC ai ai schools in trying to reseller we need to customize the rubrics as I just said and we need to provide real contact students and this thing that should be fair individual performance and we also need to think about the errors and we would like to focus on and showcase those in our programs as we need to encourage students and we also need to be a result at the beginning of the privacy and science in conclusion uh we always need the same effective we've always wanted teachers to cover from the back uh here we did well uh so we always need to visit us so we always do and it is very important for me and I'm sure it says for all the students we need the human factor we need teachers to facilitate our older people thinking too we need our teachers uh it doesn't matter what kind of tools we have we need to as teachers we need to ensure reliability and validity in terms of effect but generation to write quizzes especially the mechanical ones assess students performance provides customized learning plans for each and every student create movement they can create less than five course lines uh for example you can ask test 52 design a course design this before that course design everything you would like to start okay well as they ask teachers we don't hear ai's in our schools and generously I thought but we need to work hand in hand thank you very much for this thank you very much Miss Sheila I wanted to uh ask you a question here from our uh q a and it's from Amira perdas who says what are some examples of the inputs that students receive in class for the learning portfolio authentic assessment in your example would you like to tell us a bit about them for example if this is about black holes for that week these verticals we have two listening inputs and two uh reading books then may learn about um sorry so just may learn about black holes they can learn about the Big Bang Theory they can learn about Stephen Hawking series uh they learn about these things and then they learn about uh the lasting convention or the speaking conventions related to that input and then uh we it's like reading into writing listening into speaking I don't know if I'll answer the question well that's just yeah I think Amira wanted some examples of the inputs that they received like when you give them the assessment I think that's given it a bit more information about how you designed this estimation uh for example the question after they learn about black holes the question the writing questions can be something like this price uh how black holes are formed that could be the right form okay thank you Miss Gina uh one more question uh I have which is basically a theme of some of the questions I've seen here which is about um assessing the originality of students assessment um the assessment that they submit and someone did ask you which you touched on in your presentation someone asked how much or should we be trying to examine or assess whether the assessments we receive are original I mean should we be putting them into turning in and scanning them should we be using the different tools to see whether they've been generated with AI or or not what is your opinion on that it is important if during some speech of writing for example it is not possible for detect if it is a student's own work or just argument so for this reason it must be a very good idea it starts we do um the fact we do is from Russia University some of those tests can be written in Christ and some of those tasks can be given as homework and in that case when you see because it said portfolio you you see the color of student paragraph essays and when you see the students Chiropractic recently in class you can Paragon to the one written at home and you can very easily understand that way thank you thank you very much for your presentation uh you've got a lot of practical uh suggestions and examples first of all I want to uh thank all the speakers and just apologize that we're we're running a little over time it's five o'clock we're supposed to be wrapping up but we do uh we have received quite a lot of questions so we do want to leave a bit more time for our open dialogue as we planned so I'm just saying uh that we'll be running a bit over 5 p.m so now we can go into the open dialogue and get a couple of the questions answered but we had a lot of requests for people from people to share um presentations and things with these uh the presentations will be shared on the own Temple Merlot page we will be sharing the link with everyone and there you can um take a look at the presentations and and and maybe ask questions over there so maybe we can uh start the dialogue and I have a couple of the questions that I want to ask um I think a couple of people have brought up ethics right um ethical and unethical practices uh maybe you can and one person asks how do we how do we as Educators present um ethical modeling or how you know how to use AI ethically for our students maybe uh can I ask uh Dr Hill me to answer this question oh okay thank you Natasha uh asking me all the hard questions yeah anyway uh yeah I think yeah of course um ethically ethics are very important I think uh in that so uh how do we cater for that and how do we check for ethics I think uh of course or how do we check that the originality I think of course um now as you said just now for uh has the uh you know the AI detecting feature plagiarism you name it I think uh there will be more and more you know tools that that can offer that and definitely of course we want to check um uh how much as you know students plaguarize or you know uh use AI but again of course uh we we want to again uh cultivate the how do you say that the culture of of you know using AI ethically because if we don't understand and we just ban everything no Ai No No chat DBT so it's gonna be uh they're using it anyway so um it's it's it's our role now again changing and evolving now we have to be like invigilators of generative AI so yeah again uh hope hopefully other uh research assistants uh virtual research assistant can help us in in in in in in in you know uh checking that yeah do you maybe want to add on to that and let us know or share a little bit we've had some questions about AI detection tools where are we as far as reliability of these um and maybe yeah if you can tell us a bit more about that because I think quite a few people have asked okay um right if you're talking about the chat GPT uh some of us was wondering whether uh we could uh use some kind of tools to detect whether these are the students who are joining the originality of the students uh work uh okay we we will have to look at um in terms of I'm talking about the perspective of the technical or the technology aspect where um we um although we can see that uh chat Deputy is very powerful but yet they have a few uh limitations so we have to leverage on the limitations of the tool as well so as you can see that um although they are trying to imitate the the human human intelligence but there are some angles that the The Tool uh may not be able to uh surpass our our uh our ability our intelligence for example like uh the reasoning these are some some of the aspect that probably the education the Educators may have to look at this angle whereby in order to evaluate or to assess students students will we may have to look at some kind of some some kind of the reasoning spot and not not only the um just generating the content but probably look at the the reasoning and probably uh like uh these two they have a few limitations like for example if you look at we really look at the um some of the answers given probably if you are the expert like Dr help me mentioned we are the experts so by right we should know whether it is uh legit or not or whether it is a credible or not the the the answers given by Rich activity so uh some of the things that may not sounds well uh uh the answers given by by these two so we will have to um try to decide whether uh all these answers can be used uh or not so we will have to remind our students that it should not be it can be used as a as as the assistance but it should not be 100 taken from all these tools thank you Dr norisma so we come we can't Farm out this task to AI to do for us we have to do this manually whatever in turn it in you know there's that AI detection and turn it in which a couple of people have mentioned and also there's GPT zero what do you think about this it's not ready for us to rely on yet no okay thank you and there's also a question could we even try should we even try to how much time should we even spend on doing that what's your view on this should we spend time trying to assess whether it's AI generated in the first place you're asking too sorry I'm asking uh Dr norisma but if anyone else wants to take that question um I think I think uh Miss Sheila is trying to uh say something about this again Sheila to answer thank you very much foreign why don't you answer this question um look at the previous question I think that would be a good uh [Music] uh oh doctor says we need to create the culture uh for this reason to make students a part of it instead of dealing with uh discussing it because at the moment even we even though I turned into plugins they're not working properly and it's at the moment we don't have a tool that can detect the accuracy uh uh not the accurate but we can't really detect it but uh students get for this reason I don't think we should spend uh time on it but we should spend the time creating the culture and for you we need to make uh students a part of this maybe this would be a very good uh project student knuckles thanks Miss Sheila um uh we have a question here that leads on to whatever we've been talking about most questions are concerned about how students generating outputs from chat gbt and how to detect them somehow my question is how can teachers use chat GPT to our advantage I don't think it's necessarily wrong to use the tools available now um can I ask um to Help Me Maybe to uh talk a bit more about this yeah uh thank you for the question I think um again as I showed you the the potential of the tools um you know during my presentation yeah of course uh use it to your advantage it's it's a it's really a good uh tool to help us not only brainstorm but um to kick off to give ideas and to collaborate with students you know um as the UNESCO uh chat GPT guideline I think it's awesome so you have those 10 rules of you know assessor uh co-creator co-designer you name it and and but it's again I think it's about understanding how to use it uh well uh you know again talking about the prompts just now how do you prompt um and give the quality we want the quality output we don't just want an output we want a quality output so it's the thing about computers technology and Robotics uh and and AI you know uh sometimes it's not the desired output sometimes that's not what I want maybe the prompt is wrong maybe what you ask is wrong so it's about um taking that remember that framework that was saying to you the create framework so I think that's important as well uh for for teachers for educators for professors to look at um and and you know uh uh design based on that I think of course definitely use that but again understanding the whole ecosystem also is important um to uh regulate for an ethical uh behaviors in Academia so we've gone over time so unfortunately I will have to wrap up now I would like to thank all our speakers and uh and the academics who contributed their videos I want to thank um Dr norisma Dr helmy Miss Sheila also uh to live in and Dr Miguel uh and Dr zahid and um Jason who who are here right now I want to thank all of you for attending the session today so you if you see in chat there's a bookmark there from our own temples Merlot page oh hey um uh we can continue the discussion there we will be sharing the materials from today's session there it is past 5 PM in Malaysia right now so we're going into the weekend time uh we will share a a QR code for the feedback form we would love to hear from the participants um uh the comments about the webinar and we will be compiling all the questions and responding to you after the webinar is over requests for esets you will be able to do them in the uh in the feedback form so I'm going to close the webinar now thank you to everybody and hope to see you again soon thank you very much everybody thank you thank you bye [Music] thank you foreign [Music] thank you [Music]

Channel: UM Languages & Linguistics

Views: 1,297

Keywords:

Id: OsJtbKmJY7k

Length: 133min 5sec (7985 seconds)

Published: Fri Aug 04 2023