Speaker 1: Hi, my name is Andrew McLaren and I want to help you navigate AI and children and their privacy. So it's really important to keep students' data private and there are some pretty major concerns if we're looking at COPPA and being compliant with COPPA, which is an act in the United States which is protecting children's privacy. So we'll go over in this video what is COPPA? How, for example, AI's privacy practices in terms of services kind of line up or compare with COPPA? And some top tips for parents. So if you want to skip ahead to that part, that's fine, but I think it's important to have some context because my tips are going to be related to, like, these other things. So these slides were generated using AI as a basis. Like AI didn't make these slides, but I had a conversation with ChatGPT actually around this topic and that helped me kind of inform some of the parts of this talk. But the ideas are mine. They're just kind of structured a little bit with a little help. AI is great for use like that and we'll talk about how teachers can use AI for purposes like that and it's totally okay, but other things are not such good ideas. So COPPA is the Child's Online Privacy Protection Act from 1998. Now you don't need to read the whole document, but I did want to show you kind of like what the document is. It's pretty long. It's pretty intense. So the main points that you're going to see, and I saw this as a teacher teaching at a public school for about seven years. I was also on the tech team. So I kind of saw like the school norms and procedures around student privacy. So children who are under 13 cannot have data collected by default. That's broadly speaking, they shouldn't have data being collected without parent permission. So if there's data being collected by someone who's under 13, the parents need to give that written permission and there needs to be a way to submit that written permission so they can document that they have parent permission. The reason behind this is, you know, essentially kids, their data, they're not super careful with it and they just put stuff out there that may make them get targeted for certain ads. So if they really like Pokemon, then they might get ads for Pokemon. And that is seen as like they don't have good enough defenses against that. So targeted ads shouldn't really be targeted at specific children. You can have like a kid's ad and that's okay. But not like this kid likes this thing, so we're going to just like throw a ton of ads at them. That's okay for someone who's over 13 legally, but not under 13, which I would even say should be pushed up to higher ages, more like 18 or so. So in a lot of European countries, there's just a straight up ban on like advertisements aimed at children, like in Sweden. So this is kind of like a weaker version of that, but it's the same sort of idea. Like we want to keep kids' data safe so that they're not getting targeted by big corporations and they can't really defend themselves against that. And there's, you know, there's some other concerns too. Another one that comes to mind is they may be giving up identifiable information about themselves or like things that you might see as like security questions, like what was the first city that you lived in or your first pet, that kind of stuff. All that information could be mined from their prompts that they're putting into these AI things. So that is also stuff that we want to protect. So we want to make sure that confidential information is protected or identifiable information is protected so that that person isn't being targeted in a couple of reasons. So the main thing here is that parents need to give consent. They can only protect, or they can only collect like some information, like there's some information that they're allowed to keep and some that they shouldn't be keeping. And these are really the obligations of online services, like they have to do certain things to make sure that they are COPA compliant. Of course, not all data is going to get caught doing their procedures, but they have to have good enough procedures showing that they are taking this seriously. And this is really important for educators to follow. So if you're a parent who's choosing, do I want my account on this platform or not? This is a very different thing than if you are a teacher. Because if you can't get written consent as a teacher, you cannot use it in the classroom. It's not compliant with COPA if you're doing that. So like at the school I was at, they had a list of websites that the school used. And when parents signed up for the school, they had to sign that document. Basically that covered up most of the websites they were using, but they updated it every year. So there were some websites that some teachers used and they didn't put it on that list. I didn't even hear about this list my first year. There's some gaps and the systems aren't perfect, but there are laws around this. They have to do this. There's some serious consequences if they don't. So this is what ChatGPT says. The data they collect are the inputs that you're putting in. And if you're scoring the feedback. So that's going to be linked to your account. They also use cookies so that they know what location that person's signing in with and that type of information. What other accounts are logged in while they're logged into this. So they are collecting information from cookies here, which should be concerning immediately because it's like, are we getting to opt out of the cookies on this website? So that should be a question on any AI website. Am I getting to opt out of some of these cookies? And then if you're using a third party to provide data, like if you're signing in with Google or Microsoft, those are pretty common nowadays. A lot of websites have a sign in with Google option. That's going to give some information. So if you go and ask Google, hey, what information do you give when you sign in with Google? It's going to give name, email address, and profile picture. So the profile picture could be viewed as potentially data that you want to keep secure. The email address, a lot of school districts kind of have the location in them, or if you know the email address, you know what school district it is, so you can say what city the person's living in. So there's a lot of information that people can figure out by looking at data that's being stored here. So they say they tokenize the data, which is just another word for encryption. So they're using some sort of code to take your message on your computer and send it over to their servers. And then they're using that code to also send that back to you, but like it's jumbled up while it's sent. So that's, and they're saying that that's the data is reasonably safe here. I can get that. I'm not a computer person, so I'm not going to speak to the strength of their encryption, but the data is being collected. And as far as I can tell, there's no limit on how long they're storing it. They're just keeping all the data so they can keep on looking back at this. And I have some concerns about this. I think that like this, this seems relatively secure. It seems like they're kind of doing their due diligence, but if that data is being stored somewhere and it's linked to my account, they know that that data is linked to my account somewhere. So they do know like what information is being said for what person and that information is being stored somewhere. So it could potentially get hacked. So that's a concern for me. So this is a major issue, the age verification process. I was just reading an article from the New York Times about, they were talking more about like meta and Facebook and like social media having age verification processes that are just not good enough. They're not really keeping minors off of this. So OpenAI, they're saying that the platform is not for people under 13, but you can use it with parent permission. There's no ability for students to create an account that says under 13. When you create an account, it will just not let them create it. So there's no way for minors to submit parent permission and there's no way for minors to create an account without lying about their age. So they all know to lie about their age. That's just like a social norm at this point. Like all the kids, if they want to use something, they're just going to lie about how old they are. And this is an open secret is, you know, what we like to call us like companies all know that minors are using this. They're mining all that data and they're saying that they don't really know who's going to be a minor or not, so they're just able to mine the data because they don't know. I don't think that that's good enough. If you are using Google sign-ins, you may not realize that you're sharing certain information. And if you're using a Google account that's for someone who's under 13, does it prevent you from signing in? I bet you it doesn't. Maybe in the future, they'll fix that. But currently, if you just have a Google account, it'll just let you sign in with that Google account. And I just really don't think that they should be storing this information for minors. There's other things that I've heard students using like chat GPT as like a therapist, you know, like COVID was hard to get some mental helps. A lot of people moved towards trying to use these things to talk about their life. Not only is that potentially sensitive information in terms of their security, it's like private information that could be embarrassing and could be used for blackmail, you know, like this data that people are putting on here is very compromising potentially. So the main idea here is if you have an account with their system, they have data about you and they know your name, email, picture, plus whatever you're plugging in into there directly. And other things that their cookies may be collecting, which isn't very clear right now. So there's a lot of issues in terms of them actually, like just having minors on the platform and not really doing enough about it. And then having data that I don't think that they should have of minors as a result. So here we have my top tips for parents and educators. Just summarize them all in one slide so you can see them clearly. I personally don't think that students should be using AI directly. I don't think that's responsible. It's upsetting to me because AI is a powerful tool and I've learned a lot using it. I do want students being able to work with AI in the future. But this is one of those things I think that the cost is too great. It's not worth the benefit. And, you know, people may differ on this. So I wanted to give some advice for people who are like, okay, but I still want to have my students using AI. I still believe that this is worth it. A couple things is that you really need to coach your students about what personal and private information is and what that means in the context of these AI companies. They shouldn't be putting anything like security questions related or identifiable information. You know, the general things like don't tell strangers stuff, like don't tell AI that stuff. Like this is just a part of that digital citizen kind of responsibility. So we want to make sure that we're not signing in with school accounts. And if they are making an account, that they're just making up some information that's not their real name. It's not the real email. Some of these websites now let you use it without signing in. So that's better. So if you want to have students using a website without signing in, that's a pretty good option. In addition, students should be supervised when they're using the AI because it can tell them some pretty wild things. Some things are just outright dangerous, like telling them to do experiments which are going to hurt them or like I've seen it not be able to give CPR advice correctly. So I've seen on Google's how do I do CPR and they get the AI thing, that's bad. Really dangerous, right? It doesn't know how to make a lab safe. So it's, there's a lot of issues with this. If you want to use AI for like grading or a seating chart, it's really difficult to do that because you're going to have like students' names on there. If you remove the names from the assignments and then plug those into ChatGPT, I think that that would probably be okay. You know, it's not ideal because it's still kind of linked to you and the school and it could still be kind of figured out potentially. But I think it's pretty reasonable. If you remove the names. But if it's something that's like, hey, link to your Google Classroom and we're just going to be able to take things from there, don't use that. Like that's going to have a lot of information due to those Google accounts about all the students and everything. A big piece for me is we got to advocate for some things. Like there's right now the landscape is not where it should be in my opinion. I think we need to have some more oversight and some more hefty fines for people who violate this stuff or corporations that violate this. We need to have better sign-in requirements. That's a big piece here is that we don't just let minors walk into any private business. Like if they're going to go into like a bar or something, you need to have a valid ID. And like that's a verified document, right? So we can have some sort of similar process for some things online. Or at least a limit on how long the data is stored so that if it was to get hacked that they don't get everything from like all of history that you've done on there. So also a really important thing to keep in mind is AI's ability is changing constantly. These terms of service are also pretty fluidly changing. So make sure you're staying informed. Check how old like any source of information is on AI. Like see how old this video is. There may be some things that I'm saying in here that are still relevant and some things that are not. This is, it's kind of tangential. I would say that the copyright thing isn't so much about like student privacy, but it is related because students' work that they've made in your class can be thought of as them being the primary author and therefore having copyright. If they haven't told you that you have permission to use their student work, then you can't use it or like you can't claim it as your own and then like give it to people. So they need to give permission for it to be given to other people including this corporation. So it's somewhat a minor issue with student work. I see this a lot more in teacher groups where they're just taking images of textbooks or of lesson plans made by other teachers and they're saying like give me some multiple choice questions from this. You can't do that because you're giving like a page of a document that has so many copyright things on there, images and text and whatnot, that you don't have permission to. Like those aren't your things. So you can't just give that to someone else because now they're going to use it and they didn't pay for it. So we got to make sure that we're not uploading anything that we don't own our copyrights to. Thank you for McLearning with me. Before you go, please check out the following services that I offer with my business. So if you click the link in my videos, this goes to my website. I might be changing it from Podia because I'm no longer doing the interactive videos for sale. I might put those up for free. But most of my other services are linked here. So you can see that I've got like a Teacher's Paid Teachers and a Wise Ant. The Teacher's Paid Teachers has a lot of my assessments that I have for like the NGSS, as well as lessons that I made when I was actually a teacher in the classroom. So I've got quite a few things like CERs and other things on there. If you want to get one-on-one live support from me, either like with science or like working on like teaching support lessons and that kind of stuff, you can sign up through Wise Ant. I have a link to there where you can contact me there and we can schedule hourly appointments. I've got a few other things on here like a Facebook page and other social media. And if you're interested in getting professional development with me for like a team of teachers, a whole page with information on that. So yeah, feel free to check that out. Thank you.