Follow the show in your app. Apple leads on iPhone, Spotify on Android.
Summary
Most conversations about kids and technology are about keeping children off it. Tim Estes makes the case for the other path: technology built so well that a child can actually use it safely.
Tim Estes has spent about 25 years in artificial intelligence, from a company he started at 20 to work with the defense and intelligence community and Wall Street compliance. At the nonprofit Thorn, AI he helped build has been credited with identifying roughly 12,000 trafficked children in the US. He joins The Ryan Vet Show to argue that parents do not have to be at war with technology, but they do have to understand the business model behind everything their child touches.
After a 10-year-old named Nyla Anderson died following a viral challenge served by an algorithm, Tim set out to build a safer alternative rather than only warn parents away from screens. The result is Angel Q, an AI tool for kids he calls a perplexity for kids, which generates age-appropriate answers on the fly and runs every question through a three gate system that decides whether to answer, soften, or defer, and notifies the parent. His practical rules are blunt: keep technology in the open, never in private, and read the incentives before you hand over the device.
Key takeaways
1
The answer to kids and technology is not only keeping them off it. The other path is technology built so well that a child can use it safely by design.
2
The most useful question a parent can ask about any app is not how it works, it is how it makes its money. The business model tells you whose side the product is on.
3
Never let technology have a private relationship with your child. Keep screens in shared spaces, never behind a closed door.
4
Angel Q, which Tim Estes describes as a perplexity for kids, runs every question through 3 gates: answer, soften, or defer, and notifies the parent.
5
AI is not new to protecting children. The kind of AI Tim Estes helped build at the nonprofit Thorn has been credited with identifying roughly 12,000 trafficked children in the US.
6
Innovation and protecting kids are not in conflict. A product can be built for both at the same time.
Terms defined
Plain-language definitions for the ideas in this episode. Structured for search and AI answers.
Angel Qnoun · product
An AI tool built for children that Tim Estes calls a perplexity for kids. It generates age-appropriate answers on the fly, uses playful personas, and runs every question through a three gate safety system. It is subscription based, with no ads and no resale of children's data.
In this episode: Tim's alternative to handing a child an open search bar or an adult chatbot.
the three gatesnoun · product design
Angel Q's safety system. Every question a child asks is either answered directly, softened into an age-appropriate response, or deferred, and the parent is notified.
In this episode: How Angel Q decides what a child can see.
the velocity gapnoun · Ryan Vet framework
The widening gap between how fast technology advances and how slowly human wisdom, morality, and guardrails catch up.
In this episode: Why the tools reaching children keep outrunning the protections around them.
Chapters
Jump to any moment. Timestamps deep-link the audio.
The guest
Tim Estes
Founder, Angel Q · Chair, Alliance for a Better Future
Tim Estes is the founder of Angel Q, an AI product built for children, and the chair of the Alliance for a Better Future, a pro family group focused on human dignity and technology. He has spent about 25 years in artificial intelligence. He started his first company at 20 years old, then built systems for the defense and intelligence community and for Wall Street compliance, work that put him inside the hardest problems in the field long before the current generative wave. At the nonprofit Thorn, AI he helped build has been credited with identifying roughly 12,000 trafficked children in the United States.
His path to building for kids started with a loss. After a 10-year-old named Nyla Anderson died following a viral online challenge that an algorithm served to her, Tim decided the right response was not only to warn parents away from technology, but to build something a child could actually use safely. Angel Q is the result, an AI tool for kids that Tim describes as a perplexity for kids. It generates age-appropriate answers on the fly, uses playful personas like a sportscaster and a French chef, and runs every question through a three gate system that decides whether to answer, soften, or defer, and notifies the parent. It is subscription based, with no ads and no resale of children's data.
Tim's message to parents is not fear, it is literacy. You do not have to be at war with technology, but you do have to understand the business model behind everything your child touches. Keep technology in the open, never in private. And know that innovation and protecting kids are not in conflict, because you can build for both at the same time.
00:00Never let technology have a private relationship with your child
Tim Estes ·On this episode of the Ryan Vet Show. If you're going to let your kids use technology, they always need to be doing it in the open and never privately.
Ryan Vet ·Yep.
Tim Estes ·And so anything you can do to deny the technology having a private relationship with your child is critical.
00:24Why this conversation is different
Ryan Vet ·Welcome to another episode of the Ryan Vet Show. I'm excited to have a conversation today that's different than some of the conversations we have. A lot of our guests so far have talked about how to prevent your kids from getting on screens or accessing the internet or waiting until a certain age, which are all great tactics. But the reality is we live in a world where technology is ubiquitous and kids are going to have access to it, whether it's at school or their friend's house or even in your own home. And so I'm excited to have a guest with me, Tim Estes, today, who has an extremely interesting background spanning many different industries, but always around technology and artificial intelligence. Tim, thanks for being here.
Tim Estes ·Ryan, thanks for having me on. Really appreciate it.
Ryan Vet ·Absolutely. And you have such a fascinating story that takes you from Wall Street to defense and everything in between, often around the idea of artificial intelligence. And artificial intelligence isn't just five or six years old, from when ChatGPT became free in November of 2022. It's been around for a while. Do you mind giving us a little bit of your story, some of the projects you've worked on over the years that you're able to share? And then we'll transition to where you are today and how you're protecting kids.
01:36Tim's story: from AI at 20 to Thorn, Nyla Anderson, and Angel Q
Tim Estes ·Yeah, well, thanks, Ryan. I often joke that AI is the only thing I've ever done with myself since I graduated from college. People ask, what have you done, what are you engaged in? And they expect some rattling off of whatever got me into AI recently. And I'm like, no, I was 20 years old, third year undergrad at UVA, and I decided to build an AI company. And they're like, when was that? Well, I'm dating myself now, but it was 2000. It started because of a Bill Gates interview in Red Herring. They were talking to him about, if he would build a company today, what would it be about? And he said, well, why can't computers write their own code? There was all this money pouring into dot coms, and he said, why can't we make some progress in AI? It's a huge area. And that really intrigued me. So I took up that challenge and got pulled into a lot of things. Like many things in life, you start off with a certain vision, and while the vision might hold, the path to get there generally is not the way you expect it to play out.
Tim Estes ·We were building something that was essentially a way to make the very early social networks, the pre-Facebook ones, when they were just about posting stuff, more relevant by using AI to understand news. It got good at understanding documents with algorithms. For the time it was pretty interesting. You could throw a bunch of documents into a computer and it would tell you, this is about this world leader, or this is about this type of activity overseas. Some people in Army intelligence got intrigued that it could sort out some of the information management problems they had. We were working with them, and then 9-11 hit. And 9-11 was really a call to arms. I've never been someone who could go serve and fight. I would have been a pretty lousy soldier. So when you're a geek, it was just a way of being a geek trying to serve. And for over a decade, that was what I focused the company on. We supported multiple wars and many different activities, got to know some amazing heroes in the intelligence community and the defense department. And from that journey, we built a really amazing piece of technology. It got invested in by In-Q-Tel, which was the CIA and the intelligence community's venture capital arm. It started to proliferate into different defense areas.
Tim Estes ·Right around that time, we took it to Wall Street in the early 2010s. The investors said, we believe in your government business, but we didn't fund just a government business. So go pick at least one commercial thing to do. We decided to go at the banks. We created a way to read all the emails and chats in a lot of these banks. When someone said, big blue is going to buy this company, the algorithms before were too dumb to know that big blue equals IBM, and we could link those two with our AI and say, that's a possible insider trading issue. Think about someone working the merger and acquisition side of an investment bank throwing that line to a friend who is also trading. There was such a bonanza of misbehavior in 2007 and 2008, and all the crackdown came in 2011 through 2014, after regulations like Sarbanes-Oxley and Dodd-Frank passed. So we were putting technology into compliance to help the good guys in banks root out the few bad guys. It ended up being very successful. We got about half the major global banks on Wall Street using the technology. Some of the work went into the healthcare space here in Nashville. I'm a Nashvillian, proud native. I love building things here in the Southeast, though our funding pretty much came from Silicon Valley or New York.
Tim Estes ·But along the way, from some of the government work, a really amazing nonprofit called Thorn came knocking. This is one of the most remarkable groups on the internet trying to combat the plague of child sexual abuse. They were doing interviews with children that had been trafficked for sex online. And the interviews showed that three quarters of those kids were first advertised on the open internet. Back then it was on a version of something like Craigslist called Backpage. And this is the thing that really moved us, we cried when we heard it. When they interviewed these children that had been rescued by the police, half those kids that had posted their stuff online had been forced to write their own ad by their pimp, so it would sound young and attract pedophiles that would pay more. The women that run that, some still run Thorn today, and others have gone on to great things like the Heat Initiative. Sarah Gardner is an amazing human being doing a lot in this space on technology harms now. They were heroic, with huge hearts that could live with hearing that kind of heavy thing. And they came looking for partners that could help them make a dent on it. We were happy to be one of those.
Tim Estes ·So we built a system with them, using the AI we had at the time, this is around 2012 to 2014, where we would read all these ads and figure out that they were likely put up by underage people. The technology has been there to effectively figure out that kids are online in places they shouldn't be for a long, long time. I know, because we did it, on a harder version of that, without nearly the signal that some groups have to build from. Over the last eight or nine years since it went live, it's probably found 12,000 kids that have been trafficked online in the US. And we were just a tool provider and builder with Thorn. The law enforcement, the thousands who have used it, they're the heroes who intervened and found these kids. But it was a huge privilege to be part of that mission. And it made me very aware of the issues that can impact kids online. That's an extreme thing, kids being posted online and moved from city to city, seeing the same kids show up in different cities. Literally in the data, seeing sales of trafficked children right there. It was a staggering moment.
Tim Estes ·I got more politically engaged back then. Senator Corker, my senator from Tennessee at the time, led a lot of the hearings on child sex trafficking. We had some breakthroughs with Tim Tebow and Senator Hawley recently in the Senate on some big issues, so the fight goes on, but Corker was one of the real leaders. He brought Ashton Kutcher, who helped create Thorn, up to testify. And in one of the first moments where politics intersects with a real problem, it led to the passage of FOSTA-SESTA. That law is what made Section 230 no longer cover people who were aiding child sex trafficking. It's insane that there was a law that gave immunity to people aiding and abetting child sex traffickers, but that was the law left for us. I'm sure Senator Wyden and others did not intend that, but it was so badly written, so broad and ambiguous, that courts had interpreted it that way. It took changing the law to fix it. So I was very active trying to help there, but also busy running a company.
Tim Estes ·And this comes up later, because when that company came to its fruition around 2020 or 2021, we merged it with another amazing archiving company in the financial services area. And I finally got a brief break. As soon as I had a little time, friends of mine sent me something about a little girl. It came out of a conversation. A friend and I were talking about what our kids were seeing in schools. They were both first graders at that point. And we said, well, what's in school might be problematic, but what's online is so much worse. And then, like proof, he sends me an article about a girl named Nyla Anderson. Nyla was a 10-year-old girl in Philadelphia, went to charter school there, spoke three languages, was on TikTok. TikTok had only recently bought its way into the American market by acquiring a company called Musical.ly. Musical.ly had all the silly dance videos of under-13-year-olds, and it got in trouble, it got big COPPA fines because it shouldn't have had those kids on, it didn't follow the law. And then TikTok absorbed all that. And Nyla is proof, because she's 10 years old on TikTok. Instead of the silly dance videos that she and probably her parents thought she was getting, they serve up, in this for-you algorithm, which is their way of creating a rabbit hole to get you hooked, a video called the blackout challenge. The blackout challenge shows you a way to choke yourself, and if you don't do it right, it doesn't tell you you're going to kill yourself. And so this sweet girl attempts it in her mother's closet and suffocates. Her mother finds her and takes her to the hospital, and she passes five days later. It's so emotional to me, because I finally met her mother at an event that the nonprofit I chair, Alliance for a Better Future, ran with Sarah's group Heat and others, to bring the largest set of survivor parents ever to Washington, DC, to bring attention to these online harms and to harms from AI that are impacting kids today. I'd never met her mom until then. So that was a moment. And I'll keep telling Nyla's story until we pass the law so it never happens again.
Tim Estes ·So that got me really moved into this. I had access to talent in the AI space and access to capital. It's very hard, frankly, to build a business that serves kids in a way that's healthy for them instead of just another piece of candy. It's really easy to make AI addictive and gamified for kids. That is almost an unfair fight. But if you're trying to only use AI as a tool that's constructive, with key guardrails so it cannot foster unhealthy engagement, and you're policing that, it's not easy. I think we've made a great tool with Angel Q, the company I built from that moment. But it also requires a cultural shift, where parents say, if we're going to let our kids use this technology at all, and I mean under-13-year-olds, and honestly even 13 to 16 has challenges, then it's a gradation of freedom. With Angel Q, we invented a very solid, effective, and completely safe tool to the very best of our ability. At the same time, a parent has to recognize there's a window to have fun with technology that can be very limited, and then there's all the real world and relationships we want kids to foster, and then there's technology as a tool, which should be designed well and doesn't have to be fun. If you confuse technology as an escape with technology as a good tool, that's the beginning of the challenge. And the truth is, most parents use technology as a babysitter. Nicki and others who are friends of ours have brought this up. Some people say stay off the screens entirely because it's such a slippery slope once it's a babysitter. I don't think it has to be. It's a software problem. It's a business ethics problem that manifests in applications and how they're monetized. I think it's a terrible world where the people who care about their kids are essentially at war with technology, which is one of the great things unleashing human productivity and potential. That feels like a terrible trade. And I want to provide, and probably many others do too, technology that's so aligned with the parents and the family that you can use it as it's meant to be used, as the tool, and put it away when you don't need it.
17:20The alternative to abstinence
Ryan Vet ·Yeah. I think that's so good, Tim. You truly have been a part of technology at its earliest roots. Your story about hearing Bill Gates talk about AI and programming itself is fascinating. I just saw a clip today of Steve Jobs at a dinner talking about something similar, what if you could talk to a computer and have a real-time conversation about anything, talking to philosophers like Socrates and Aristotle. And that was late 90s or early 2000s. So it's been around a while, and you've been a part of it. You are in a group that isn't anti-technology, and I appreciate that, because that's been the easy answer. If it's bad, you need to abstain. But you're saying there's an alternative, a happy medium, because technology provides new value and unlocks human potential. So let's talk more specifically about Angel Q, what you've created that lets kids use technology in a safe, helpful, educational way that also gets parents involved. It brings it back to the TV in the family living room, no longer a device locked away under the covers.
18:46What Angel Q is: a perplexity for kids
Tim Estes ·Yeah. When we thought about the problem, which is that kids get access to applications that are an ecosystem unto themselves, that's always the problem with social. Angel Q is almost orthogonal to social. There's no overlap, no competition, because there's not even the ability to friend someone on it. There's no communication between people on Angel Q. Instead, it's much more like a souped-up search engine tool that's using AI. The closest analog is a really good solution for search called Perplexity. So Angel Q, in a lot of ways, is like a Perplexity for kids. The AI goes out, and there are different layers. You can chat with it in a general way, but it's bounded, it won't keep a conversation going indefinitely. It will follow through a few times and then try to nudge you off. It's not trying to sustain a long session. It's there to stimulate the kid's personality. And it doesn't just work by chatting, it's a talking thing. We have these little personas, almost like talking to cartoon characters. The personas are limited. They know they're an AI. They never make up some story about what they've done or where they've lived, because that's all fiction. But when you have a sportscaster talking about what happened in an MLS game on Saturday, it's kind of fun.
20:29Fun personas and the French chef
Ryan Vet ·Yep.
Tim Estes ·And we wanted to make this kind of fun. We have a French chef, and having the French chef talk about cooking is actually a good party trick.
Ryan Vet ·Yeah.
Tim Estes ·And I watch adults do this all the time. One of the little secrets of Angel Q is half the parents or more say, hey, can I use this thing? Because it's actually a really good explainer. You ask it a question, it talks back to you, it highlights the words so that kids who are early in learning can almost follow the bouncing dot, and it brings back images that put the story in context. That's a normal engagement. And then if you want to go deeper, and we found in our own household this has been really key, if your kid's doing a research project, we had our son do a project on Glacier National Park. He wanted to dig in, and it could not do the work for him, but it could take the challenge of, how is he going to find out about this? Are you going to sit next to him at the computer while he goes on Google? What if Glacier is the name of some person somewhere whose picture we don't want to show up? Parents know what I'm saying. When you hit that search bar, you're spinning the barrel of that gun of overexposure, and you don't know exactly what's going to come up.
21:57The problem with a search bar for kids
Ryan Vet ·Yeah.
Tim Estes ·And I've had a lot of parents, and I say this self-confessionally, I don't necessarily have a kids' profile version of my web browser and search engine. So a lot of times they're coming in, I'm with them, and I'm the one putting it in. Bad stuff doesn't come up that often, but the search bar doesn't know to stop it.
22:16Age-appropriate answers on the fly
Ryan Vet ·Right.
Tim Estes ·Even if I have it defaulted to safe mode for myself, it's still not going to hold the same standards for what a 6-year-old should see versus a 10-year-old or a 12-year-old. Angel Q is pretty unique. It translates the content to be age-appropriate on the fly. So if you're searching for something and there's a Wikipedia article on it, and Angel Q is doing the research and that's one of the top hits, when it brings that back, it has translated it into what your 6-year-old needs if the child is that young. And with anyone younger, and I've said this on Nicki's show as well, anyone eight or nine or below, you should do the tech with them. There's even a case, given some of the unknowns in AI, that the whole 13-and-below experience is one you should be present for. As they get into the eight-to-13 range, having them use an iPad on the kitchen table while you're cooking dinner is fine. But if it's a younger kid, you're probably right there over their shoulder, in fact probably the one typing or talking. Even with younger kids, and by the way, a four-year-old, I wouldn't want them on that screen at all. Once you go below eight, below being a reader who can consume the language, the parent is really using this as a tool to explain something to the kid. That's probably its primary value, an aid for the parent to explain something. And then once they start being asked to do their own research, which in school is generally between seven and nine, there's a window to do that alongside the kid. But there are so many guardrails in this thing. Content-wise, we've had 40,000-plus conversations, and we've not had anything explicit or problematic get flagged the whole way through. Now, whenever you have a track record that's perfect or near perfect, you know it won't last forever, so I'll say that as a caveat. I'm sure there are things we've missed, we're human. But we are ruthlessly diligent on safety in terms of toxicity. In some areas we may have even over-calibrated, like conversations that can go into violent subjects. There's probably a way to communicate something like war safely to a child.
25:03The three gates: answer, soften, or defer
Ryan Vet ·Right.
Tim Estes ·But we've been pretty calibrated. Younger kids can't really ask those questions without it hitting a deferral. So one of the things Angel Q does is, when a kid asks a question, the system looks at it as, is this a question I can actually answer for a kid of that age? It decides if it can, and then it translates any answer to be appropriate to that age. The second gate is, that's close, that's a gray area. In that case it says, I'm going to give you a little bit of an answer, and I'm going to translate it even younger than you are, just to be safe, and I'm going to notify the parent in the background. And the third one is, nope, that kid shouldn't be dealing with that, that's a defer topic. The thing comes back and deflects. It'll say, this would be good to talk to a trusted adult about, how about we talk about something you could do to help your three-point shooting, if it's the sportscaster. It deflects, and it immediately, in real time, notifies the parent that the question was asked. Every conversation the kid has is fully logged. It's a browser, essentially, so it can bring in content from YouTube through their API, and from Disney Plus and Netflix, and render it in our kids' browser. That content has a system behind it where we can screen what it can access. There's a whole parental ratings component where a parent can say, on a scale of one to four, I only want a one on any violence, or a two on any dark themes. And we can preview and show you the kinds of things on Netflix and Disney Plus that get screened out when you do that. So Angel Q has a whole bundle of layers to make sure a kid's browser-like experience is driven within the parents' values and with awareness. It was a way to say, there's a whole lot of stuff on the internet that could be useful, healthy, and fun, and we can winnow down to just those things. It's a better middle ground than nothing, or hoping for the best. We've also built learning modules. We found a way to take content and YouTube videos and turn them into lessons where kids can learn about giraffes or about Pokemon. And it's not just that they watch it for entertainment, they're actually quizzed on it, so it builds a habit of engaging, remembering, and learning. So we're trying to turn Angel Q into something that's truly positive, healthy screen time, while still bounding it substantially. And this is backed up by our model: it's a fairly inexpensive subscription. There's no advertising, no resale of data, nothing like that. We have no way to monetize this except parents sponsoring it.
28:46The business model: no ads, no data resale
Ryan Vet ·Yeah.
Tim Estes ·And that's by design. It's getting easier for us over time, because the technologies we shipped a year ago are cheaper to deliver today and higher quality. We were pretty early. There are other kids' browsers and kids' AI products going out there now, I've seen three or four. But we've got a year-plus of in-the-field proof of its safety, and three years of development and design, including one of the only truly open-source safety frameworks, called KidRails, that we built with a leading American AI company. We put that out there so people can look at it, make their own decisions, and potentially build from it. The other piece with Angel Q is that, as a company, we're starting to have conversations with other large kids' companies and tech companies, because they're realizing that before they go into kids, the responsible ones want to be careful. A lot of the toy companies have been very responsible. There are a lot of app companies that have been pushing AI at kids with reckless abandon, but if you look at the big brands, the ones I've talked to have a conviction that they do not want to put a kid at risk until they understand more. They won't even roll the dice on it. One very popular deal that came out was Mattel working with OpenAI. A lot of us were concerned about getting close to a leading model vendor and putting that out while the safety guardrails weren't in place. To their credit, they probably used it to help design new ideas, but they didn't put it out interactively. I think a lot of these companies have great ideas for cool experiences with their IP, but they haven't figured out how to be comfortable with the safety part. So we at Angel Q think we've figured out part of that, and we're now working with them, saying, we can empower you to have AI-powered experiences and tooling for these brands with deep assurance that it's going to be healthy for the kid and not harmful.
31:26Weaponized AI versus AI for good
Ryan Vet ·That's so good. And Tim, I appreciate what you've done. You have a unique perspective, seeing so many different facets of AI from the early days, and using it now for good. Where a lot of people, and you hinted at it, but I'll say it more bluntly, there are a lot of companies out there who have used AI and weaponized it to get attention, to create addiction, to build their own businesses. Business ethics is a conversation we could talk about for hours when it comes to AI. But as we wind down, I'd love to ask this final question. What can parents do today to be safe with technology, not to remove it? Besides Angel Q, which is a great solution I think they should pursue, what are some practical tips from things you've observed and seen?
32:24Practical tips: keep technology in the open
Tim Estes ·Well, there's a lot that can be done in your household that's just good cultural practice. Our kids are younger, both not teenagers yet, so this is probably easier for us to enforce, but I would say that if you're going to let your kids use technology, they always need to be doing it in the open and never privately. Anything you can do to deny the technology having a private relationship with your child is critical. Almost all the horror stories that come out are the kid going to their room, getting into something innocently, getting led down a path, and terrible things happening. And sometimes terrible things happen that don't take their life or cause immediate harm, but they feel ashamed about what started, and they get more tied in that way. The only way to stop that is to keep it in the open. Now, you can't prevent what kids may be exposed to at a friend's house. So I almost wish, I would love to talk to the Bark people about this, that you could use geospatial positioning on the devices to allow an open mode when they're in the home and a very closed mode when they're not. That would be pretty cool.
Ryan Vet ·That'd be a great feature.
33:57Bark mode on steroids
Tim Estes ·I don't know if anybody's done that, but if I'm a parent, the thing that would make me more open to giving a device with true internet connectivity would be knowing it's essentially in Bark mode on steroids when they leave the house, or Gabb mode. There are some really good companies in this space doing ennobling work. I don't know that we have that exact thing yet. If they come on your show and say, no, we've got it this way, then hats off to them. For now, my kids have access to iPads that are time-limited and locked down, and they have to use them in public, not in private. It's sad, because just like with social networks, the idea that kids could chat with other kids online and connect isn't necessarily bad. It didn't have to be bad. It's probably more dangerous now than it's worth, so don't do it. But the real issue is that when the social networks grew organically, everyone was signing up while the internet was young, but when they had to keep growing revenues at a rate different from their growth in users, that made the incentive about increasing time spent online. And when you align time spent as your value, you can tell yourself people are spending more time so you're providing value. But a great way to keep people hooked is addiction. There's no better path than to addict people. So the kids were told they could go to the playground and meet other kids, and the people who owned the playground turned it into a casino. They let adults into the playground, and they put the bar right next to it. As a parent, I would never send my kids into that playground. There are whole ecosystems like that, gaming sites I can name, all built that way, because the people who created them weren't morally moved, when the first kids got hurt, to say, shut it down, let's figure it out. Half a decade or a decade later, that has never been their conscience. So you really do have to vet it. If your kids are using technology, figure out who makes it and how it's set up. You don't have to be a scientist, but everything they're touching has some business model tied to it. If it's a video game and I'm paying for it, well, how much time should they spend on it, and is it overstimulating? If it's YouTube, how much time are you spending, and are they getting fed a bunch of short videos that make their attention far worse and drive dopamine desensitization? Maybe that's something you should do together. What do they need to watch on YouTube that you can't put on the big screen and watch together? As a parent, thinking about the use case and the business model tends to reveal it. And when the business model isn't aligned, you ask, what can go wrong? And the answer is, it only has to go wrong once. There's no mulligan for what happens to your kids.
38:06Understand the business model behind every tool
Ryan Vet ·That's so good. I had Tom LeNoble on the show, who was early at Facebook and did a lot of their compliance, and shared a lot of what they had to see. He talked about the many subpoenas he got, and how he actually helped, before his departure, identify cases of kids who were being exploited on the platform. But I think you said something so profound that I want to repeat, Tim, the idea that business is tied to just about everything. There's nothing free. So we have to understand the business model driving whatever platform or tool we use. That's such a profound and easy thing for parents. You said it well. You don't need to be a scientist. You don't have to pick apart the code. It's simple: how are they making money off your time or your attention? Well, Tim, I want to thank you so much for this incredible conversation. How can people find out more about you, about Angel Q, and start putting these practices into place today?
39:18Where to find Tim and Angel Q
Tim Estes ·Yeah, well, thank you. You can find Angel Q online at the App Store. We run on iOS. We haven't gotten onto the Google platforms yet, we will at some point, but we're still refining it. You can also read about us at angelq.ai, that's our website. You can find me on LinkedIn under Tim Estes, and Estes AI is kind of my little thing there. And I'm also on X, where you'll find me engaging in a lot of these issues. Something else I'm deeply involved with, a second hat, is a group I co-founded with several other mainly conservative, center-right, pro-family groups, called Alliance for a Better Future, which I chair. We have an amazing CEO, Janet Kelly, who was Governor Youngkin's head of Health and Human Resources in Virginia, a cabinet secretary twice, and one of the leaders in getting phone-free schools done in Virginia. She and our team, and our chief officer Ari Kagan, who has been engaged on AI safety for some time, are fighting to show that it's more important to put human dignity first. Much like the Pope's recent statement about human dignity and technology, our leaders need to support human dignity ahead of expediency by these companies. And it won't lead to less profitable companies if the profits are legitimate.
Ryan Vet ·Right.
Tim Estes ·But when you have digital narcotics companies powered by AI, which unfortunately I think there are too many of, those are talented people. Everyone targeting ads and using AI to manipulate children, they're not curing cancer. They're not doing the use cases with AI that were sold as the great future hope. So the government, by allowing this to continue, is subsidizing bad uses of AI. There are many pro-family groups, from the Institute for Family Studies to Heritage, on our council. We believe fundamentally that innovation and protecting human beings are not in conflict. What's in conflict with the family are digital narcotics and exploitation business models, and cheap profits. If someone were giving political cover to Pablo Escobar, I wouldn't cheer Pablo Escobar as an innovator. And we now have the digital versions of that going on.
Ryan Vet ·That's deep. That's profound. Well, Tim, I appreciate the work that you're doing, your presence on X and LinkedIn, and everything you're doing with Angel Q. It's so important for families, and I appreciate that the message isn't just a business solution for you, but something you're passionate about and pursuing to make change with policy and everything else. So, Tim, thank you so much for being on the Ryan Vet Show today. And everyone listening, thank you again for listening to this episode. Until next time, inspire forward.
Tim Estes ·Thanks for tuning in to the Ryan Vet Show. Be sure to subscribe, comment, and like this episode. Plus, share it with someone who needs to hear it.
Frequently asked
Is keeping kids off technology the only way to keep them safe online?+
No. Tim Estes argues abstinence is not the only answer. The alternative is technology built well enough that a child can use it safely by design, with guardrails that decide what to answer, soften, or defer. Parents do not have to be at war with technology, but they do need to understand how each product makes its money.
What is Angel Q?+
Angel Q is an AI tool built for children that Tim Estes describes as a perplexity for kids. It generates age-appropriate answers on the fly, uses playful personas like a sportscaster and a French chef, and runs every question through a 3 gate safety system. It is subscription based, with no ads and no resale of children's data.
How does Angel Q decide what a child can see?+
Angel Q runs each question through 3 gates: answer, soften, or defer. It answers safe questions directly, softens sensitive ones into age-appropriate responses and notifies the parent, and defers questions it should not handle while alerting the parent. Across more than 40,000 conversations, the team reports no explicit or problematic content getting through.
How should parents evaluate the apps their kids use?+
Ask how the product makes its money. Tim Estes says the most useful question is not about the code, it is about the business model behind everything a child touches. If a product profits from attention or data, its incentives run against the child. If it earns money another way, like a subscription with no ads, the incentives can line up with the child's safety.
Why does Tim Estes say kids should use technology in the open, never in private?+
Because a private relationship between a child and a device removes the oversight that keeps children safe. Keeping screens in shared spaces lets parents stay part of the conversation. Tim's guiding rule is simple: never let technology have a private relationship with your child.
What is Tim Estes's background in AI and child safety?+
Tim Estes has worked in AI for about 25 years, starting a company at 20 and later building systems for the defense and intelligence community and Wall Street compliance. At the nonprofit Thorn, AI he helped build has been credited with identifying roughly 12,000 trafficked children in the US. That work shaped why he built Angel Q.
Who was Nyla Anderson, and why does her story matter here?+
Nyla Anderson was a 10-year-old whose death was linked to a viral online challenge served by an algorithm. Meeting her family moved Tim Estes to build a safer alternative rather than only warn against technology. The lesson he draws is about accountability for how products are designed, and the case for tools that protect children by default.
Content warning: this episode discusses online harms to children, including child sexual abuse, trafficking, and the death of a child linked to a viral online challenge. Listener discretion is advised. To report suspected online exploitation of a child in the US, contact the DHS Know2Protect Tipline at 1-833-591-5669 or the CyberTipline at report.cybertip.org.
Most conversations about kids and technology are about keeping children off it. Tim Estes wants to talk about the other path: technology built so well that a child can actually use it safely. He joins host Ryan Vet to walk through about 25 years in AI, from a company he started at 20 to work with the defense and intelligence community, Wall Street compliance, and the nonprofit Thorn, where AI he helped build has been credited with identifying roughly 12,000 trafficked children in the US.
His path to building for kids started with a loss. After a 10-year-old named Nyla Anderson died following a viral challenge that an algorithm served to her, Tim decided the right response was not only to warn parents away from technology, but to build something a child could actually use safely. That became Angel Q, an AI tool for kids he calls a perplexity for kids, with age-appropriate answers generated on the fly and a three gate system that decides whether to answer, soften, or defer, and notifies the parent.
Tim’s message to parents is not fear, it is literacy. You do not have to be at war with technology, but you do have to understand the business model behind everything your child touches. Keep technology in the open, never in private. And know that innovation and protecting kids are not in conflict. You can build for both at the same time.
In this episode:
Why the answer to kids and technology is not only abstinence, and what a safer alternative looks like
How 25 years in AI, from defense and Wall Street to hunting online child trafficking with Thorn, shaped Tim’s approach
The story of Nyla Anderson, and the moment Tim decided to build Angel Q
What Angel Q is, and the guardrails behind what Tim calls a perplexity for kids
The three gates, answer, soften, or defer, and how a child’s question gets handled and a parent gets notified
Why every app your child uses has a business model, and how to read it
Why kids should use technology in the open, never in private
A hopeful case that innovation and protecting kids are not in conflict
