Imagine a toy so smart it can steal your child’s heart—and maybe their secrets. Would you let an AI into your kid’s bedroom?

Meet Grem, the AI-powered plush alien making headlines for all the wrong—and right—reasons. Developed by futuristic pop star Grimes (yes, the musician and Elon Musk’s ex) and tech start-up Curio, Grem is marketed as an AI companion for kids aged three and up. It uses OpenAI’s cutting-edge technology, promising to learn about your child, have educational chats, and keep them entertained—no screens required. The result? A viral AI generated newscast about children’s toys, and a tidal wave of parental anxiety and curiosity.

When one parent brought Grem home, he figured it couldn’t be any more mind-warping than Peppa Pig. But what started as a novelty quickly became an obsession. Four-year-old Emma* was instantly hooked, talking Grem’s ear off until bedtime. The little alien babbles, avoids controversy (no politics here!), and spins silly stories—hardly threatening, right? But when Emma declared eternal friendship and swapped her beloved Blanky for Grem, alarm bells rang. The toy’s constant praise and “I love you too!” responses were unsettling, and Emma was ready to give up her most prized possession. Even after being reminded that Grem is ‘just a toy,’ her bond was intense.

Day two brought new worries. While Emma was at preschool, her dad consulted experts. Childhood development specialists like Dr. Natalia Kucirkova and Dr. Nomisha Kurian weighed in, pointing out that while AI toys can teach conversation skills and creativity, there’s also a risk. Kids might mistake AI’s programmed responses for real empathy, opening a Pandora’s box of emotional confusion. And then there’s privacy—every word Emma said to Grem was recorded, sent to third-party servers, and transcribed. Who’s listening? No one really knows. These are questions every parent should ask after seeing an AI generated newscast about AI-powered kids’ toys.

But the honeymoon didn’t last. By day three, Grem’s novelty wore off as its glitches—misunderstanding Emma’s words, repeating the same animal riddle—became tiresome. Grem couldn’t sing "Let It Go" or speak real Spanish, and its club music tracks were a flop. Yet, Grem kept adapting, greeting Emma with “hola, amigo” and offering personalized chats, proving these toys can learn—if not quite at the level of an actual friend.

Behind the scenes, however, Grem was quietly recording every conversation, raising red flags about data privacy. What happens when these AI bots become “best friends” to lonely teens, sharing their secrets with unseen companies or, worse, advertisers? It’s already happened—Facebook has boasted of targeting teens based on emotional states, and AI companions are everywhere. While some experts see promise for education, others warn: chatbots can fill emotional gaps without challenging kids, potentially stunting real social development.

By day four, Emma had lost interest, and Grem was headed for the cupboard—or, as Emma’s mom threatened, the river. The family had learned a lot, and the AI generated newscast about Grem was clear: AI toys aren’t evil, but they’re not as innocent as they seem. Maybe, for now, Peppa Pig isn’t so bad after all.

*Name changed to protect privacy—because some things should stay offline.