AI Toys Are All the Rage, But Are They Safe for Kids? Here Are Some of the Options Our Experts Recommend

AI has been creeping into our daily lives slowly but surely (or quickly, depending on your perception of time). It’s seeping into our lives in ways most of us are not even fully aware of. One of the biggest such instances? Toys.

With the holiday season upon us, holiday shopping looks a little different than it has in past years. AI integration is now making its way into children’s toys, and there’s a lot to unpack—especially considering we’re still, technically, in a somewhat rudimentary stage of artificial intelligence. Which begs the question: are AI toys safe?

Plenty of factors are at play, and the answer is nuanced; there’s still a lot of gray area. One of the biggest risks is exposing children to inappropriate content. Some toys—particularly companion devices—may discuss mature topics and explicit material. In the same vein, they also run a risk of creating unhealthy relationships with your child, which could have long-term effects on their social skills down the line.

Our resident toy expert, Cat Bowen, still has fears about the viability of AI integration in children’s toys. She says, though, that in a closed, highly regulated system, you may be able to proceed with caution.

“Mostly, I don’t trust them unless they can show it is a closed system and that there are an incredible number of prompts blocked,” Bowen told me. “But for things like storytelling and anything animated, I mostly don’t trust it because too many things have gotten through and become very scary.”

The perfect example is the Miko Robot. Developers were pressured to add an AI “off switch” after a data leak led to a website where visitors could download thousands of responses that Miko generated to children. Some of these interactions were highly personal and in some cases, even included specific children’s names.

We wouldn’t recommend companion toys or anything that includes data or privacy concerns—ie. anything with a camera. At least not yet. One device we’ve tested, though, checks a lot of boxes for us—the Stickerbox.

The Stickerbox is one of the biggest toy announcements of 2026. It generates stickers on command, using a voice-to-image prompt, and prints them in real time with thermal printing. Your child can see all of their magical ideas come to life in seconds. Like ChatGPT, for example, they can iterate on their ideas to get the final result they want.

The all-in-one station doesn’t match the exact closed system Bowen mentioned, but it comes close. For a true closed system to work, interactions must be strictly limited. During her tests, she did get it to generate an unsavory image, but it took several tries and very specific prompting. With the correct oversight, it can be a viable option. The keyword here is oversight: parental discretion is a must.

The Stickerbox also includes built-in protections to keep your child’s ideas wholesome; they’d really have to go above and beyond to get it off script. It also features screen-free interactive play—always a plus in today’s screen-dominated environment. The device also boasts enhanced encryption to protect your family’s data, and we love that it uses BPA-free and no-BPS printing paper for its stickers.

While we don’t recommend most AI-integrated kids' toys on the market, smart toys are a solid alternative at least at this stage. They don’t have full-on sentience; they’re more reactive and still require human interaction. Think of them as a simple technological upgrade for interactivity, rather than an AI-driven overhaul. The new Smart Bricks, for example, react to each other like an ecosystem and require certain sequences to activate, rather than actual human prompts. For example, children physically turn on the smart brick to activate smart play. To get the intended result, they’ll sometimes have to shake the object or place a minifigure in a certain spot.

We included some of our favorite smart toys below, most of which have been editor-vetted and tested in some capacity.