Tech in EdTech

The Implications of Gen AI on Data: A Conversation with 1EdTech

Magic EdTech Season 1 Episode 51

 In this episode, Beatriz Arnillas, Vice President of Product Management at 1EdTech, and Parag Samarth, Chief Strategy Officer at Magic EdTech, discuss the TrustEd Apps initiative by 1EdTech. They discuss the program's evolution from focusing on data privacy to encompassing security, accessibility, and a new GenAI data rubric. Beatriz explains the importance of transparency, organizational preparedness, and data literacy in using AI in education. The conversation highlights the need for critical thinking and responsible data use to create safer and more effective educational technology environments. She also encourages educators and edtech providers to work together towards a safer, more mature, and reliable industry.





00:00.00

Parag Samarth

Firstly good morning Beatrice! This is Parag Samarth. On behalf of both, yourself and myself and Magic and 1EdTech, I wanted to you know greet the audience and the listeners. So good day to you all and we look forward to a fairly interesting and exciting conversation with 1EdTech. As an introduction I'm Parag Samarth, I am the Chief Strategy Officer at Magic EdTech. I am also responsible for the technology initiatives at Magic. Personally, I am very passionate about data and what it can do in the educational field, and especially with the huge focus on AI within that, the LLM and Gen AI. I think that the whole issue of data has become very very critical and important. And, therefore not surprising that today we would be talking about GenAI and its implications on data. I would like to introduce our fellow participant on this podcast webinar Beatriz from 1EdTech.


01:28.70

Beatriz Arnillas

Good morning Parag. Thank you! I'm Beatriz Arnillas and I'm the Vice President of Product Management at 1EdTech, previously known as IMS Global. I'm happy to be here this morning. A little bit about myself, I started my career as a teacher in Peru and then I moved to the US, where I got my Master of Fine Arts and I held the position as an art professor in a university for several years. I got involved in technology because I was interested in technology as a media for the arts. And then soon discovered the potential for teaching and learning via blended and online courses. So I moved from teaching and learning to higher education administration and K12 leadership to introduce technology in the Houston public school system where we did a very large implementation back in 2013. So while I was doing that in Houston we discovered the standard specifications from the then called IMS Global. You know things like common cartridge and things like that. And so that's how I ended up working for now called 1EdTech and just working to provide education with standard specifications that make technology connections easier and help move data around the learning ecosystem.


03:04.21

Parag Samarth

Right. Interesting combination of art and technology. In fact, I was told all along that the best technology people in the field of computers are people who come from an arts background. So I guess you made the right move. So let me get into the topic of the conversation today. This is about the initiative that 1EdTech has championed for quite some time called the TrustEd Apps and I'll have you explain to us what that means and what that stands for. And how the TrustEd Apps initiative has now evolved with the availability or in the presence of GenAI and how is the new TrustEd Apps initiative especially from a GenAI perspective looking like. So firstly if I could request Beatriz. If you could explain to us what exactly is the TrustEd Apps initiative and how does that morphs into the I would say upgraded, updated TrustEd Apps initiative which focuses on the GenAI data as well.


04:34.33

Beatriz Arnillas

Right, of course. The TrustEd Apps program started of course initially with concerns about data privacy when it comes to the use of technology particularly in K-12 education, but in education in general. Because of course, the education technology products have access to a large amount of data and there were always concerns about who owns the data. Who is the data shared with? And making sure that we only share the learners and the user's data with the people that must have the data. So the concept of data minimization which is common here in the US as well as in Europe. So the TrustEd Apps program in 1EdTech started with the data privacy rubric and certification. But more recently that program has evolved to include the security practices specification and rubric, the accessibility specification and rubric, and more recently the Generative AI data rubric, which we have developed to a point and now have called for the participation of a larger group to develop it more completely so to speak. But the idea is that in the field of education, it's our responsibility to create an environment that is safe, private; where only the people who should have access to data, should have access to data but it also addresses concerns about who owns what data. So that's in broad brush strokes so to speak.


06:12.90

Parag Samarth

Yeah, now I think that's interesting, that you know very rightly you start with the initial focus on privacy. But now you have added more pillars - security and accessibility and the GenAI. From an industry perspective you know we've been all initially very confused, then very excited, and now I would presume very realistic about the implication of GenAI especially from an education perspective. And while on one hand there is a lot of interest, a lot of excitement, and a lot of people are saying what are you doing about it? On the other hand, there is a lot of concern about - Am I inadvertently using somebody else's data? Is my data going into somebody else without my knowledge? By you know, getting into this, how am I exposing my you know IP and my privacy issues and all that? So I think from that perspective having a specific pillar for GenAI makes a lot of sense. Because whether it is privacy or security I think it has a whole new meaning from a GenAI perspective. So if you could take a minute and explain to us how the GenAI pillar in the TrustEd Apps - what specifically does it address? And how is it implemented and administered?


07:40.83

Beatriz Arnillas

Yes, of course. So first of all, there are 2 fundamental or maybe 3 fundamental actions or programs I think institutions should be concerned with. The first one is that there has to be an organizational preparedness. You know where I'm talking about leadership values, policies procurement, practices, communications, involving students, and all users in data governance and controls including maybe audits and professional development et cetera. And of course including the concept of AI and data in the curriculum and preparing the learners for the future with AI. So for that 1EdTech has developed it along with our members primarily really with contribution from our members an AI preparedness checklist which is meant for the leadership of an organization. And then on top of that, we have developed the TrustEd Apps generative AI data rubric. The instrument we've developed is that starting point. It's by no means completed. It's the first step with the questions that we're encouraging users to ask of edtech providers is for example, the edtech providers should warn the users that AI is being in use if it's being in use.  Is there options for opt-in or opt-out - I'm thinking here of student age and things like that or even of student preparedness. So if they're not prepared yet to address the use of AI perhaps, there should be an option to opt out. Whether the AI tool is a third-party product or an internal tool - and if it's a third-party product - of course there's all kinds of other data concerns to ask. Then the source of data to build a corpus of the tool -  where does the data come from, how valid is that data of course, and an explanation of how the AI data model is improved over time. That's the first few questions we ask in our current rubric but we have made a call for participation from our members and we've had a tremendous response I should say, to continue to develop this rubric to add additional concerns related to data when AI is in use, right?


10:10.47

Parag Samarth

Right. I think you mentioned about the source of data, and I was just kind of relating to the fact that increasingly now, all the social media platforms and browsers are making available GenAI capabilities on their platforms. I won't take names but you know everybody knows that use any of these standard search browsers and they give you the option to use GenAI. Interestingly, when I asked certain questions which, you know like what's my address and you know what's my date of birth? It came back and said we don't share private information, which kind of made me feel more comfortable. So there is a self-regulated I, would say, practice amongst these platforms to protect privacy that is good. And then when I asked questions which were not so private, which was you know like what's going to happen in this situation and so on and so forth and I was getting a fair amount of response. At the bottom, I was also getting a bibliography of sorts saying this is where the data came from, which again was quite comforting that I know that this is being sourced using public data and not somebody's private data. So you're right I think we are having those things taken care of. Some of those questions as to how is the privacy and security handled. Where is the source of data? I think that is consistent and I do believe that there are some federal regulations that are also coming in. I might be a bit out of touch over here. But which I believe requires that personal information not be shared and most importantly, the source of information should be made explicitly available. So I think from that perspective it's an initiative from your side but it's also now becoming probably mandatory from a federal regulations perspective. So to summarize where we are as of now you know, we heard about the TrustEd Apps initiative. We heard about how it has morphed into additional areas of pillars including privacy, and security and accessibility and now the GenAI. And the fact that you know we have the 5 questions, but just to be sure and clear these are self-assessment questions for the platform provider or for the content provider to kind of assess themselves and this is not a kind of a certification. Is that correct?


02:28.55

Beatriz Arnillas

Right. The privacy rubric is a certification. We do certify companies on privacy. But the security practices, the accessibility rubric, and the AI - Generative AI data rubric which we're still in the process of building are self-attestation instruments, and the idea is to provide both suppliers and users information about the questions we should be asking or answering to provide the education field as a whole an opportunity to grow and mature in its practices as we all learn more about…security is a constantly evolving field as you know accessibility I think we still all have a lot. Like our task force said, accessibility is a journey, not a destination.


03:21.26

parag Samarth

Right.


03:23.60

Beatriz Arnillas

Right? So it's in the spirit of continually learning and evolving for the better good. And so we're very lucky our members are both suppliers and education leaders and together they help us build these instruments.


03:33.78

Parag Samarth

So you really get both sides of the table so to say - suppliers saying here is how I'm going to certify myself and users coming back and saying how is this going to be of help to me and use to me right? I mean that way you're kind of getting a meeting of minds and in the event somebody does have a concern saying you're missing something then, your platform would allow that conversation to happen and those concerns to be addressed.


04:08.70

Beatriz Arnillas

That's right. That's the intention to build transparency, promote dialogue and to provide that opportunity to exchange because I'm sure you know this - the brightest technologies doesn't always understand all the nuances in education and vice versa.


04:22.92

Parag Samarth

Yes, yes.


04:26.52

Beatriz Arnillas

So by having both voices in the room we're solving, we're bridging this gap that is very difficult to bridge sometimes.


04:33.93

Parag Samarth

Right, right. Absolutely. And also you know there is a strong balance required and a tradeoff required between how can I make things so much more easier and so much more interoperable and all that. But at what cost? And even a small child going to kindergarten, we have to respect his or her privacy and you know not not share information which might at some point of time be detrimental to the person and as we know a lot of this information tends to hang around in the social media so you know it's important to take these precautions right from the beginning and not necessarily later on. So I think from that perspective it's an important aspect. Do you have any examples of how this GenAI data rubric has been used in a real-life situation? So you know if some of our listeners either from the supplier or from the user side are wanting to know what has been the experience of people? What have been the follow-on questions that people have asked? Is there something that you can share with us?


05:49.14

Beatriz Arnillas

So the first thing I would say is that even the process of inviting some of our supplier experts to talk to us about Generative AI data and asking them - interviewing them - asking them questions, clarifying questions.


06:02.32

Parag Samarth

Hmm.


06:08.16

Beatriz Arnillas

Even that process in order to build the rubric has been a tremendous opportunity to learn for all of us involved. So far the rubric was just published in January and we're in the process of publishing it in what we call our TrustEd Apps management suite which is…


06:19.95

Parag Samarth

Okay.


06:28.23

Beatriz Arnillas 

…Available to all our district members and university members. The idea is to do exactly what is happening right now which is people are asking questions - “Okay I didn't understand, I didn't think about that, that you're right.” People are beginning to talk about what's bias. What are hallucinations? What questions should I be asking about the source of data for the model? Who owns the data that is produced by the AI when it's in use in my institution? Just people having those conversations is what our intention is to help people develop an awareness for the questions and the conversations they should be having about AI. But of course, it can't stop there, can it? There are many more questions, for example, teachers, I was talking with a high school teacher during the…


07:18.36

Parag Samarth

Hmm.


07:25.46

Beatriz Arnillas

…recent conference and he was explaining to me that he teaches history. He said I know my students are using AI to write their papers in my course and so it…we are even having conversations about what, how should a teacher design their lessons knowing that their students might be using AI to write the papers. Our intention is that the students learn not that AI learns.


07:49.20

Parag Samarth

Yes, yes.


07:55.36

Beatriz Arnillas

So these conversations and this dialogue are bringing to the forefront the need to develop not only institution-wide policies and procedures but also what are the pedagogical concerns. How are the teachers to think about their lessons if your students are using AI to write papers?


08:11.40

Parag Samarth

Correct.


08:13.52

Beatriz Arnillas

Shouldn't you ask them to write a paper from using 3 different tools maybe and comparing the results, then doing a fact check, and returning and analyzing the results, etc? So teaching data literacy including AI data literacy to the students because like you mentioned they are using social media and they're putting all their data out there are they aware of the implications? So yeah.


08:34.88

Parag Samarth

Yes, exactly. Yeah yeah, and I think just like you know some of us have gone through various stages of learning. I remember when I was in school and we didn't have computers, you know the emphasis was also how well you write in terms of you know your handwriting and so on so forth, and that kind of in a way directed the way I was assessed as a student. And then later on as people started having access to computers, emphasis increasingly you know became where people submitted their responses online, and therefore the emphasis shifted from the handwriting to what response was given. And then kind of raise the bar in terms of saying now that you have more time to think about it, you need to be more critical in your thinking, you need to give opinions about what you think. And I think from that same perspective if GenAI is going to create a lot of content, I think you made a good point saying “Okay, what are the 3 different platforms saying about the same topic and you as a student - firstly, what do you think and you know, what is your critical thinking? So it's…I think if properly used it's going to push the bar higher and it's probably going to ask for a much more deeper level of critical thinking. And for students to frame their thought process you know and I think it's kind of going to get them to train their minds for a lot more critical thinking and analytics than what was the case earlier. I think that's probably where I see this going, right?


10:43.80

Beatriz Arnillas

I agree with you. So there's no question that these tools are going to save time. We know already of teachers that are building lesson plans using AI, so that's good. They're saving time. The next thing is to make sure that the users of AI are aware that the bias is still a problem even if you train your system or counter the bias. There's still some issues with bias and accuracy and of course, the issue of intellectual property hasn't been sorted out yet right? That will have to be sorted out soon.


11:06.73

Parag Samarth

Right.


11:18.81

Parag Samarth

Um, yeah.


11:23.80

Beatriz Arnillas

But essentially any information that is out there in the web is liable to be captured by AI and if some of it is the property of an organization or an institution is that okay? We haven't solved those issues yet, have we?


11:37.90

Parag Samarth

Right, right. And I think you mentioned bias which is a very important aspect. And even to notice the bias requires a fair amount of critical thinking and analytics in the sense that you know, you have to look at the outcomes and then you have to compare the outcomes over different platforms over periods of time. And then you have to be able to spot that bias which itself is a huge capability. And you know inadvertently I would say that GenAI has kind of made that point, that part extremely important and I now see a lot of people spending fair amount of time trying to spot biases and then understanding where that bias came from. So that itself is a skill by itself which is crucial.


12:31.70

Beatriz Arnillas

Yeah, yes I was trying to find some information about how to counter bias and there are several methods to counter bias, but sometimes when you try to counter bias it kind of goes and has bias in a different or opposite direction. So we haven't solved the problem.


12:44.93

Parag Samarth

Yeah.


12:50.16

Beatriz Arnillas

So what's important to know is this is great. It could be a great tool but we have to train ourselves to fact check and to ask more questions.


12:58.89

Parag Samarth

Yeah, so I think probably as you said this was launched in January so it's early days. It's just about three months since it's been around. The fact that we are talking about it already obviously means that it is gaining a lot of attention and traction which is good. I would believe that over the next three-four months we will start seeing the outcome of how this is performing in the marketplace. What kind of responses are coming from suppliers, are the users satisfied with what they're getting? It's going to, as you rightly said, lead to more questions. Maybe 5 questions will become 10 questions as we go forward and hopefully at some point of time some kind of a certification capability. So, that we just don't take the user or the supplier self-assessment but we're able to confirm to ourselves. And I increasingly expect to see this condition becoming part of the RFPs that we will see from school districts and higher ed institutions which will then make it so much more formal. So I think that that is a good input to our listeners. As we wind up, any last comments, anything that you want to leave us with some thoughts? So that you know we can have the users, listeners ponder about it and then hopefully when we come back and do the next round we might have responses from them and we will definitely see you coming up and telling us where this whole initiative has progressed and how it has matured?


14:50.57

Beatriz Arnillas

Yes, I'm very proud of 1EdTech. I think 1EdTech fosters a community of experts both from edtech companies and from educational institutions including higher ed and K-12 leadership.


14:56.14

Parag Samarth

Um, yes.


15:05.66

Beatriz Arnillas

All these groups work together in task forces, project groups, product steering committees to inform what it is that the field of education needs. I encourage educators and edtech providers alike to join us, to join our ranks. We're really making a difference in education. These rubrics are designed to support our collective journey towards a better edtech. Safer, more mature and reliable conditions, where we provide teachers with better time, more quality time to prepare and we provide differentiation and choices and accessibility to the learners. So we can address learner variability more effectively.


15:52.71

Parag Samarth

Right.


15:56.26

Beatriz Arnillas

So we welcome everybody and we hope you will join us in this path. Working together we definitely find better solutions.


16:04.99

Parag Samarth

Right, right. So that's important and I fully agree and echo your point that 1EdTech is fulfilling a very important role in the education industry. You know some of the initiatives - early initiatives especially regarding standards are now becoming almost commonplace. You know if you don't tick the box then you're not in business. And you know I've also seen how those standards have evolved as we as you see new challenges coming up. And I've seen those standards evolving. So I think that kind of gives me the same confidence in the organization and the initiatives you're taking. So let me end this conversation on this note. Thank you again Beatriz for giving your inputs and your insights. Like I said you know, you being a teacher yourself and now a technology professional, I think you bring both sides of the table together which is quite interesting. For our listeners you know in case you want to know more about 1EdTech, I encourage you to visit their website. It's 1edtech.org and you'll be able to see all the initiatives and work streams that 1EdTech is championing. All the events that are coming up. And hopefully, some more of these discussions that we will have between ourselves from the industry and professionals from 1EdTech. So thank you again and wish you a good day. Thank you, Beatriz.


17:50.17

Beatriz Arnillas

Thank you, Parag! Thanks for having me here.