By SCN Staff

Quick Review

• An estimated 22 million AI-integrated toys were sold globally in 2025, including toys marketed for education and companionship.
• Common Sense Media found that approximately 27% of tested AI toy responses were not appropriate for children.
• Avoid AI companion toys for children under 5, and use extreme caution with children ages 6 to 12.
• Before buying, investigate the AI model, training data, storage, third-party sharing, safety testing, parental controls, and company shutdown plans.
• Look for meaningful transparency, not just “smart,” “educational,” or “kid-friendly” marketing.

The toy aisle is changing quickly. A stuffed animal or robot may now listen, generate responses, remember details, and encourage your child to keep talking. In 2025, an estimated 22 million AI-integrated toys were sold worldwide, according to a Journal of Medical Internet Research report.

That rapid growth arrived before child-development research, independent testing, and clear privacy standards fully caught up. Common Sense Media’s January 2026 assessment found that 27% of tested AI toy responses included inappropriate material, including references to self-harm, drugs, and risky behavior. The organization recommends avoiding AI companion toys for children ages 5 and under and exercising extreme caution for children ages 6 to 12.

This is not a reason to panic. It is a reason to shop carefully.

Why This Number Matters

AI toys are not simply traditional toys with extra sounds. Many include microphones, cloud-connected software, companion apps, and conversational systems supplied by companies other than the toy manufacturer.

A toy may collect voice recordings, transcripts, behavioral information, and, in some cases, facial or emotional data. A child may also share personal information more freely with a toy that presents itself as a friend.

The U.S. PIRG Education Fund’s AI toy investigation found that tested products varied widely in their content safeguards, data practices, listening mechanisms, and parental controls. Meanwhile, senators have asked federal regulators to examine AI toy safety and independent testing.

Before purchasing, treat the product like a connected device, not an ordinary plush toy.

How To Buy An AI Toy Safely

Use this checklist before placing an order. If the manufacturer cannot answer these questions clearly, consider choosing a simpler toy.

• Identify The AI Model Maker. Ask which company supplies the model powering the toy. Is it a model created specifically for children, or a general-purpose system adapted for a toy? The toy company should also explain what training data or content sources shaped the model and what limitations apply.

  Why It Matters: A child-focused product may still rely on technology designed primarily for adults.

• Find Out Where Data Is Stored. Determine whether conversations are processed on the device, in the manufacturer’s cloud, or through third-party servers for speech recognition, moderation, analytics, or storage.

  Why It Matters: More companies and servers involved can mean more places where sensitive information may be accessed, retained, or exposed.

• Ask Whether Data Trains Models. Check whether recordings, transcripts, images, or usage data are used to improve the company’s AI or shared with technology partners. Look for a clear promise that children’s data will not be used for advertising or model training without appropriate consent.

  Why It Matters: “Improving the service” can involve broader data use than parents expect.

• Look For Independent Testing. Search for documented testing of privacy, cybersecurity, content safeguards, and child-appropriate design. Seals such as kidSAFE or participation in a COPPA safe harbor program may provide useful information, but no seal should be treated as a complete safety guarantee.

  Why It Matters: A company’s own claims are not the same as transparent, independent evaluation.

• Check For Physical Privacy Controls. Prefer a toy with a physical off switch, microphone kill switch, push-to-talk button, airplane mode, or genuine offline functionality. Confirm what “off” actually means.

  Why It Matters: Sleep mode is not always the same as stopping recording or network activity.

• Review And Delete Recordings. Confirm whether parents can view recordings and transcripts, download them, delete them, and request deletion of associated accounts or backups.

  Why It Matters: Families should have practical control over a child’s digital history.

• Examine Filters And Parental Controls. Look for age-appropriate content filters, parent-managed settings, conversation summaries or transcripts, blocked topics, usage limits, break reminders, and alerts for concerning interactions.

  Why It Matters: Controls should work on the toy itself, not only inside a companion app or paid subscription.

• Plan For Company Shutdowns. Read the terms for information about discontinued services, refunds, data deletion, software support, and offline use if the company closes or stops maintaining its servers.

  Why It Matters: A connected toy can become unusable when the company disappears, leaving families with lost money and a disrupted attachment.

For foundational questions about connectivity, microphones, deletion, and offline play, SCN’s earlier AI toy guide may also help. This checklist goes further by examining the companies, models, and long-term data lifecycle behind the product.

Red Flags To Notice

Pause before buying if a product:

• Calls itself your child’s “best friend” while encouraging prolonged conversations.
• Does not name the AI model or technology partners.
• Uses vague phrases such as “may share data with partners.”
• Has no physical microphone or power control.
• Requires a subscription for basic parental safeguards.
• Makes educational or developmental claims without independent evidence.
• Provides no clear plan for deletion, breaches, or company closure.

Consider The SCN-Certified Company Program

Parents can also examine whether a brand participates in the SCN-Certified Company membership program. The program gives organizations an opportunity to demonstrate a proactive commitment to children’s safety, responsible technology, and informed practices.

Membership is not a substitute for reading a product’s privacy policy or confirming its controls. However, it can help families identify companies that publicly align themselves with child-safety leadership and ongoing responsibility.

Frequently Asked Questions

Are AI Toys Safe For Young Children?

Safety depends on the design, data practices, content safeguards, and supervision. Common Sense Media recommends avoiding AI companion toys for children under 5 and using extreme caution for ages 6 to 12.

Does COPPA Make An AI Toy Safe?

No. The FTC’s COPPA guidance addresses certain children’s privacy obligations, including parental consent. Compliance does not guarantee strong content filters, cybersecurity, developmental design, or reliable parental controls.

Should Parents Choose Offline Toys?

When practical, yes. A toy that works without Wi-Fi generally reduces the amount of information sent to outside servers. Still, review whether the device stores data locally and whether recordings can be deleted.

What Is The Safest First Step?

Before buying, find the privacy policy, model disclosure, data-sharing explanation, and parental-control instructions. If those details are difficult to locate, choose a product with greater transparency.

Build Safer Digital Habits

AI may become part of children’s play, learning, and everyday lives. Parents do not need to reject every new tool, but they do deserve clear information before bringing one home. Compare the product’s usefulness with its data collection, emotional design, safety testing, and long-term support.

For practical resources that help families navigate emerging technology, explore SCN’s parenting resources. Then join the SCN newsletter for informed guidance on child safety, artificial intelligence, privacy, and responsible digital protection.