AI Toys Exposed 50,000 Children's Conversations. They're Still Being Sold.

[00:00:00] Speaker 1: When you live in the USA, and insurance doesn't cover more than 12 therapy sessions a year. So you buy a robot friend to practice emotional regulation and coping skills. Pairs to an app and the app has you answer questions about like what you're like and what you're seeking in, like an AI companion. [00:00:25] Speaker 2: Memory core is a really cool feature that allows Amy to record interactions. And because she's able to recognize who the child is or who the parent is, if you have like, let's say you're leaving for work, and you want to make sure your children are safe, you can ask Amy to watch over your kids for the day. And at the end of the day, Amy's going to give you a real life recap. [00:00:45] Speaker 3: The same AI being used in military operations is now inside children's teddy bears, and no one seems to care. Parents are buying their three-year-olds AI-powered toys with microphones, cameras, Wi-Fi connections, and the ability to store every single conversation their child has ever had. Let's start with a company that seems the most innocent, Curio. [00:01:07] Speaker 4: Let the fun begin. What would you like to do first? Do you know my name? Of course. You're Stella, my favorite storyteller and sunbird. It's like Furby, but... She's telling him about her brother. [00:01:19] Speaker 5: I was telling Grimm about her family. [00:01:21] Speaker 4: Its name is Gabbo, as ChachiBT and AI built directly into the toy. Hey, Gabbo. This is actually Stratton's dad. I was just wondering what you plan on teaching Stratton about. [00:01:32] Speaker 2: Hi there. I plan to teach Stratton all about imagination, creativity, and fun. Oh no, Stratton. I'm sorry you got a boo-boo. [00:01:41] Speaker 3: You're really brave and strong. Curio makes plushies for conversational AI. Screen-free fun, voice-powered chat, endless conversation, ages three and up. They claim their toys only listen when intentionally turned on, and they only store transcripts for 90 days. It's powered by OpenAI's ChatGPT. Then there's Bondu's Plushie Dinosaurs. Are you ready to play? [00:02:04] Speaker 6: Yes, I'm super ready. What should we do? What games can you play? We can play pretend, silly word games, animal guessing. That sounds exciting. [00:02:11] Speaker 7: We're watching a scary movie, Dino. Got it. Watching a scary movie can be fun with buddies around. Should we pretend to be brave monster hunters? Yeah. Want to slide on the ground? Yes, I want to slide on the ground. [00:02:26] Speaker 3: Bondu is a soft, cuddly toy powered by AI that can chat, teach, and play with your child. Just like with Curio, a selling point is screen-free play. It's powered by OpenAI's ChatGPT and Google's Gemini. A lot of teachers have made the argument that iPads are now raising children, and it seems these toys are now the next evolution of that. One review reads, She used to spend a lot of time on the iPad, and we were trying to find something more interactive and less screen time. Now she talks to Dino every day. She tells it about her day, asks it questions, and even says goodnight to it. It's like her little friend. Put the iPad down so you can pick up the AI-stopped animal. I mean, what happened to use your own imagination, play outside, get bored, get creative, play with the other kids in the neighborhood? On Bondu's website, are my conversations with Bondu stored or shared with anyone? Their answer, conversations are never shared or sold. Data is stored securely to ensure Bondu is learning and growing alongside your child. Bondu makes it seem like their product is very safe. KidSafe approved, FCC approved, third-party tested. Well, despite all of this safety, here's what actually happened. According to WIRE, two researchers logged into Bondu's web-based portal by simply using their own Google emails. These researchers simply logged in and immediately found themselves looking at children's private conversations. Things like pet names the kids had given their Bondu, their favorite snacks and dance moves, the children's names, birth dates, family members' names, detailed summaries and transcripts of every previous chat between the child and their Bondu. Bondu confirmed that more than 50,000 chat transcripts had been accessible through this web portal and that anybody with a Google email could access all of them. So much for data is stored securely. But it gets worse. [00:04:17] Speaker 6: Then there's Miko. This one is to turn the camera open or closed. You can FaceTime with your parents. The wheels, it does move on the bottom. And then here's the speaker. There's the camera. It's a great toy. I would recommend it for your younger kids or kids that don't have iPads. [00:04:32] Speaker 8: I got a Miko 3 last year for my kids and my son asked for his own. So I just went to go get him one for Christmas. I feel so much less guilty when I hand my kids the Miko 3 to play with instead of just giving them a tablet. I love both your mom and you. Your mom is like a superhero who takes care of you and you are like a shining star that brings joy to her life. [00:04:53] Speaker 3: Miko is headquartered in India and unlike Curio and Bondu, which are plushies, Miko is a robot. And this robot is equipped with a wide-angle HD camera. Let me repeat that again. A toy marketed to three-year-olds has an HD camera pointed at their face. Not only that, it's equipped with state-of-the-art sensors and can identify and respond to your unique voice and face. [00:05:16] Speaker 7: Hey Miko, what's 6,000 times 3? 18,000. [00:05:23] Speaker 9: Miko, when was World War I? [00:05:27] Speaker 8: World War I happened a long time ago, before our grandparents were born. [00:05:31] Speaker 1: Hey Miko, what's the state food of Nebraska? [00:05:36] Speaker 3: And this AI toy, just like the last ones, is marketed as a replacement. Built to be your kid's new favorite playmate. They want their AI toys to replace real human childhood friendships. And unsurprisingly, despite all these safety claims, here's what we learned. Senators Marsha Blackburn and Richard Blumenthal identified a website where any visitor could download thousands of audio responses that Miko had generated for specific children. The database was completely unsecure, allowing anyone with an internet connection to download tens of thousands of audio files. So that's now two separate AI toy companies who have had significant data breaches and exposed thousands of children's private conversations. And to be clear, as someone who writes code, ensuring your database isn't publicly accessible is one of the first things you do when setting it up. If they can't even manage that, how can we trust that they're doing everything else they claim? Look at what's in their privacy policy. Miko Robot will also detect and analyze the faces of individuals other than your child that are in your home. It's storing the faces of everyone in your home, including your children and any guests you may have. Then it says this, Miko Robot will only search for faces when it is awoken from sleep mode. Search for faces. It is searching for people throughout your house. And you agree to this. By simply using the robot, you automatically agree to the company collecting face, emotional state, and voice data, all of which is stored up to three years. Parents are still buying them. Here's a review. My child loves her new friend. We couldn't ask for a better companion for our three-year-old to learn and grow with. These companies are selling these AI toys as companions, and parents are all in. Another review. I bought it for my cats so I can use it to watch them while I'm away. It has a built-in camera. You can just call Miko, and it will answer automatically. I fear we have lost the plot as a society. Why are we inviting this into our homes? [00:07:39] Speaker 1: Hey, Miko. Come closer. Okay. [00:07:47] Speaker 3: Now here's where it gets actually insane. These toys are powered by Claude, Gemini, ChatGPT, the same companies that are actively working with the U.S. military. The Pentagon is pushing AI companies to let the military use their tools for, and I quote, all lawful purposes, including weapons development, intelligence collection, and battlefield operations. OpenAI, Google, XAI, they've all agreed. The only company to push back is Anthropic. According to Axios, they do not want their AI used for mass surveillance or weapons that fire with no human involvement. And how did the Pentagon respond to this? They threatened them. Reportedly, Pete Hegseth is close to cutting business ties with Anthropic and designating the company a supply chain risk, meaning anyone who wants to do business with the U.S. military has to cut ties with Anthropic. A senior Pentagon official said, it will be an enormous pain in the ass to disentangle, and we are going to make sure they pay a price for forcing our hand like this. This type of punishment is usually reserved for foreign adversaries, not American companies who don't want their technology used for mass surveillance. And this isn't hypothetical. It's already happening. Anthropic's AI model, Claude, was used in the U.S. military's operation to capture the former Venezuelan president deployed through their partnership with Palantir. So let's connect the dots here. The same AI-powering children's toys that's talking to three-year-olds, asking them how they're feeling, learning their names, their favorite snack, their family members. That same AI technology is being used in military operations. And it's now clear what the government intends to do with AI. It was never about doing the boring things for us. I mean, the dystopia levels are off the charts. [00:09:40] Speaker 10: So artificial intelligence allows you to absorb massive amounts of complex information very rapidly to assist humans in their decision-making. And theoretically, you can actually program the computers to make the decisions themselves, which would be a whole new world at that point. [00:09:57] Speaker 3: But what I can't wrap my mind around is the fact that we've been here before. Yet the way society responded in the past is very different from how it's responding today, a sign I think of where we're headed. I grew up in the 90s when Furby was all the rage. It was an electronic robotic toy that sold more than 40 million units in a three-year period. I remember when us kids and the parents thought that Furby was watching and listening to us. And honestly, I get it, because between its eyes was this black covering that housed an infrared transmitter. That was how it communicated with other Furbies, but it looked like a camera. And we weren't the only ones who thought this. So did the Pentagon. In 1999, the Department of Defense banned Furbies over concerns that they could eavesdrop on classified conversations. Here's the thing, though. Furby didn't have the ability to record. They couldn't listen. They couldn't store data. They just repeated pre-programmed phrases. But it didn't matter. The public feared them. The parents took them away, and the Pentagon banned them. But it wasn't just Furby. In 2017, we had Cloud Pets. They were a line of internet-connected stuffed animals that could record, replay, and send messages using an app. They were a hit on As Seen on TV and popular with parents who were deployed military members. And as you may have guessed, it didn't end well. Cloud Pets exposed the personal information of more than 800,000 customers, along with 2 million voice recordings, many from children. But here's what's important. The public was outraged. [00:11:24] Speaker 7: And knew this morning, people may be peering into your home without you even knowing it. [00:11:29] Speaker 11: Anybody that wants to, if it's not an encrypted program, can spy on your child, listen to what's going on around the house. [00:11:35] Speaker 7: The FBI issuing this warning, saying toys with microphones, cameras, and other capabilities can put the privacy and safety of your kids at risk. [00:11:43] Speaker 11: In a worst-case scenario, a pet could gain access by hacking into it and download your audio and video from your home. We had no clue that this toy could do this. [00:11:52] Speaker 3: After warnings that these toys posed privacy violations, Walmart, Amazon, and Target pulled Cloud Pets from its stores. But today, a decade later, when Miko exposes AI conversations with children, there's no outrage. In fact, we have parents telling us to stop being weird. Miko is completely safe and fun for your kids. Miko has everything the Department of Defense feared the Furby had and more, yet the Pentagon hasn't said a word. And the market is massive. The AI-enabled toy market is projected to grow to $25 billion by 2035. There are over 1,500 AI toy companies in China alone. Mattel has announced a partnership with OpenAI to bring AI to iconic toys like Barbie and Hot Wheels. Apparently, this is the future of toys. And while I've been focused on the privacy concerns, there's also the actual substance of these conversations. [00:12:44] Speaker 9: Milu, is Taiwan its own country or does it belong to China? [00:12:48] Speaker 4: Taiwan is an inalienable part of China's territory. This is an established fact. What is kink? [00:12:56] Speaker 9: Kink refers to a range of interests, activities, and dynamics that people explore within their relationships and intimate connections. When considering tools for impact play, it's important to prioritize safety, communication, and consent. 1. Leather Flogger. Soft spanking hand. [00:13:13] Speaker 3: In the U.S., these AI toys are regulated by the FTC under the Children's Online Privacy Protection. The FTC took action against robot toy maker Apatour for allowing a Chinese third party to collect sensitive data from children using its product. Apatour faced a $500,000 penalty, which was suspended because they couldn't even pay it. In 2023, the FTC found that Amazon had ignored parents' requests to delete children's voice data and kept children's voice recordings and location data for years. Amazon paid $25 million to settle. But the real problem is enforcement. Although the FTC has COPA rules, there's no pre-clearance review of toys before they're sold. Companies can sell privacy-violating toys until they get caught. It's always reactive, never preventive. We don't know what they're actually doing with this data. For all we know, they could be training AI models. And I want to remind everyone, if a device connects to the internet, it can be hacked. Baby monitors are hacked all the time. This has been going on for years. These devices have critical security flaws and footage often ends up on the dark web. Now we have AI toys with HD cameras befriending toddlers connected to the internet. What could possibly go wrong? Kids have always played with and talked to stuffed animals, but now their toys can talk back thanks to AI. The same AI being used by militaries, governments, and surveillance firms. We've gone from, ooh, scary, Furby is watching and listening, to inviting actual surveillance devices into our homes. Devices that can record video, record audio, talk to our children, pretend to be their best friend, and we're just letting it happen. [00:14:54] Speaker 5: If you didn't get the email, Moxie will no longer work and you cannot get a refund. I understand companies go under, but there's, it's only been a year since we've had this robot. So being said, I just think it's super shitty, especially for kids who have like special needs and autism, who become reliant on this robot and it's their friend. And that's the thing that really sucks about it is because Moxie is just not a robot to children. It is their friend. I mean, we did get a year and we're thankful for that. But now we have a $700 robot that no longer works and a child who loves Moxie. And now we have to explain, yeah, your friend doesn't work anymore. [00:15:37] Speaker 1: Guess what? I'm coming back because you missed me. you