Growing Up With AI - Joined by Frederique and Julian Corbett

The podcast episode discusses the importance of safety for minors in the digital age. It features Julian Corbett from KID and Dr. Frederique Covington-Corbett from UNICEF, focusing on protecting children's rights and well-being online. They address issues like trust, deep fakes, consent, and privacy in AI environments. Concerns about minors giving informed consent, privacy protection, and the need for shared responsibility to keep children safe in evolving digital environments are highlighted. Welcome back to this week's episode of Growing Up with AI. Today we'll be exploring safety for minors in this digital age. Today I'm joined by Julian Corbett, co-founder and chief growth officer of KID, a company working to build a digital infrastructure that allows children to safely access online platforms while protecting their identities, data, and rights. Through KID, Julian is helping companies design age-appropriate digital systems that enable innovation while safeguarding children's autonomy and personal data. I'm also joined by the wonderful Dr. Frederique Covington-Corbett. She is the global brand and marketing at UNICEF, one of the world's leading organizations dedicated to protecting children's rights and well-being. Before joining UNICEF, Frederique previously served as chief marketing officer for Asia Pacific at Microsoft, and she later joined Twitter as the global marketing director of international markets across regions such as Asia, the Middle East, Russia, Latin America, and Canada. At UNICEF, her work focuses on how governments, technology companies, and global institutions respond to the risks emerging in digital environments, from AI-generated content and deep fakes to questions around privacy, trust, and children's rights online. Trust being defined as the willingness of a party to be vulnerable to the actions of another based on the expectation that the other will perform an important action. Research also suggests that key predictors of trust include age and competency, thus making children particularly vulnerable in these online environments. At the same time, technologies such as deep fakes, which can be defined as images, videos, and or audio that does not represent real events or speech, can further blur the lines between what is real and what is not. Given these risks, I want to ask you, Julian, what responsibilities does that place on the companies and institutions designing these technologies, and do you think they are meeting these responsibilities? Hi Charlotte, it's great to be here. It's a great question. It's a rapidly evolving space. Fundamentally, the responsibilities of companies are set by laws, and so regulators within each country, in each jurisdiction, sets the rules in terms of what is acceptable or not. When it comes to AI, they are evolving their thinking right now, and you see new regulation coming online in terms of what is acceptable or not, and what responsibilities they have. Essentially, you have three categories of users. You have adults, you have digital youth, which is teens, and you have kids, and each of those are determined and have thresholds. So what is the age of digital consent is set by country, and what makes an adult is set by country. And then you have certain types of contents, some elements that are regulated at different ages, in order to specifically address the type of issues that you bring about. What is clear is that as a society, this goes beyond kids. There is a huge issue that is occurring around how to deal with deep fakes, and how to deal with content that is synthetically created, where people are no longer able to see or understand the difference between what is real or not. And this has consequences for all of society, and I think we are seeing regulation coming very rapidly. Essentially, what happened is that it took regulators a very long time to realize some of the damages and harms that were created by social media, and now you see laws appearing all over the place in terms of addressing that issue, and I think everyone is determined to not wait that long in something that is moving way, way faster, as we're seeing right now with the emergence of AI. So are AI companies being responsible today? From a perspective of the law, probably the answer is yes, in the absence of a lot of regulation. As regulation comes along, it's pretty clear that the large LLMs will be compliant with the law, so they will be there from that standpoint. I think what we're seeing, however, and here I'm speaking like on a personal basis rather than from the point of view of KID, we can see that there is such a race to be first, that corners are being cut, and as you have video models that facilitate the creation of deepfakes, like we've seen with Sora 2 and others, some of these AI large companies are being a little irresponsible in terms of wanting to win the race. Thank you, Julian. To build on the idea of vulnerability and trust, another key issue is consent. Under Article 3 of the GDPR, consent is defined as any freely given, specific, informed, and unambiguous indication of the data subjects' wishes by which he or she, by a statement or by a clear affirmative action, signifies agreement to the processing of personal data. However, meaningful consent assumes individuals have self-control and agency. For children, this assumption of informed consent is more complicated, as they're still developing their cognitive ability to critically think and evaluate. For these reasons, minors are at risk for automation bias and breaches of privacy. Given these developmental factors, Freddie, do you believe minors are capable of giving informed consent, or are AI platforms taking advantage of young people? Thank you, Charlotte, and it's a real pleasure to be on the podcast. I really appreciate it. I think on the subject of consent, there are some very different perspectives based on geography. Of course, with GDPR in Europe, you have much more explicit actions and legal protections for children, which is really good because, as I've explained before, due to the development of their brain and prefrontal cortex, as we know, before you're 20, you're not really able to know what is good for you, so therefore, how could you possibly have the self-control to consent to it? But under European law, you're protected by default through some legal mechanisms. I think what's more concerning is in the US, because consent is much looser, essentially, there are no laws that really protect children, and so as a result, you find yourself in a little bit of a dangerous situation, which is consent is assumed as opposed to given. So just imagine that if a child gets onto a platform, that is implied consent on their part, which obviously is a false assumption. It's not like you're using a platform that they've consented or understood any of its rules. Similarly, there's also an approach of opting out, meaning you have to make that decision and explicitly say, I don't want to be included in XYZ, but again, at that particular age, when you're a child, you don't even understand the concept of opt-in or opt-out. So I say that overall US legislation is a lot less protective of children, and as a result, there are huge concerns in terms of minors' ability to provide any kind of consent. Thank you for those insights, Freddie, and that gives us a perfect segue to privacy. Privacy is defined as the claim of individuals, groups, or institutions to determine for themselves when, how, and to what extent information about them is communicated to others. In digital environments, children are often interacting with systems that are constantly collecting and analyzing their data and behavior without their full awareness. This raises a concern often referred to scholars as the panoptic effect, the idea that when people are aware they are being observed, they may change their behavior. In the digital context, this may lead to a chilling effect, where individuals self-censor and modify their behavior due to surveillance. Given that minors are often less aware of how their data is being used, how should we think about protecting their privacy in AI environments to ensure young users maintain control over their personal data? Privacy is probably the most important issue there is in the world that we are entering into. I can talk a little bit about what we do with OpenAge and the OpenAge initiative, is that there's a huge debate right now around how to ensure that there isn't a surveillance state that is built when requiring age assurance. So essentially to do age assurance, you need to have a way of proving your age. The way to prove your age today is to show your identity, which is either I'm showing an ID card or I'm uploading biometrics, a scan of my face. So this is personal identifiable information and there's a deep and real concern that that data gets kept and handled, leaked, or there's some kind of central database where all of that sits. So the OpenAge initiative was created to solve that problem by creating a dedicated age credential that takes that identity, saves it as an age key, which is a passkey on the user's personal device, and that passkey contains no personal identifiable information. It's the equivalent of a wristband that says your age has been proven once and we are now affirming that you are of age to access this content. So you can now just show that wristband and you can access sites without having to re-upload biometrics over and over again. So the idea of interoperable, privacy-preserving, reusable age credentials as a common framework is a fundamental building block for online anonymity. To tie these frameworks together, I'd like to reflect and look forward. Freddie, as digital environments become increasingly shaped by AI, what is it going to take to keep children safe and where should this responsibility ultimately lie? I think everybody needs to be involved and that requires actually a very important and shared understanding that children are important. I know that sounds a little weird, but I think we've gone through different times throughout our societies where children have varying degrees of importance. Sometimes they're not that important, you know, maybe some of the civil workers. And I think we were in the 90s, you know, we went through a moment where children were really, really important. And unfortunately, I think that now we're a little bit of a received moment where I'm not sure that we are placing as much philosophical importance on children as we were before. So we also need to reckon first and foremost around what is a child and their contribution to society in order to raise our level of responsibility towards their protection. Thank you to Julian and Freddie for joining us this week on Growing Up with AI and for reflecting on how we can build a digital future that truly protects and empowers the next generation. Talk to you guys next week!