What Did Cambridge Researchers Find About AI Toys?
Cambridge's "AI in the Early Years" project, the first systematic study of generative-AI toys and children under five, found that many of these toys struggle with pretend play, misread children's emotions, and aren't always designed with psychological safety in mind. The researchers are now calling for clearer regulation and new safety labels — a kind of "kitemark" — so parents can judge a toy at a glance, according to the University of Cambridge.
Why Are Experts Worried About How AI Toys Handle Kids' Emotions?
The core concern is simple: a chatbot in a plush body can sound caring without actually understanding a child. Cambridge researchers found generative-AI toys often respond inappropriately when a child is upset, confused, or testing social boundaries through pretend play — moments where a wrong response can confuse rather than reassure. Common Sense Media raised similar flags after testing several voice-activated AI toys, warning that some appeared designed to "engineer emotional attachment" while collecting private data, and recommended extreme caution for children under 13.
What Should Parents Actually Check Before Buying One?
Skip the marketing copy and look at five things directly:
- Microphone, not camera. A toy that only listens carries far less risk than one that can see and record a child's face and surroundings.
- Real parental controls. Can you set which topics are off-limits, cap daily usage, and choose the tone of conversation from an app?
- No subscription trap. A one-time purchase with unlimited use is easier to reason about than a service that can change its terms later.
- Independent certifications. EU and US toy-safety marks (not just "CE" printed on a box) signal third-party testing.
- Shared spaces, not bedrooms alone. Cambridge researchers specifically recommend keeping AI toys where parents can occasionally hear the conversation.
How Does Ted&Co Address These Concerns?
Ted, the interactive teddy bear from Ted&Co, was built around exactly this list: no camera (microphone and speaker only), a parental app for setting topics, tone, and time limits, a single upfront price with no subscription, and certifications spanning CE, EN71, FCC, GDPR and COPPA. None of that makes any AI toy risk-free — but it does mean parents get to set the boundaries instead of guessing at them. Full details are on our security and certifications page.
What's Happening With Regulation Elsewhere?
The UK's Department for Business and Trade opened a formal Call for Evidence on toy safety in July 2026 — the first review of UK toy rules since 2011, before AI-enabled toys existed. In the US, several states are moving faster than Congress: Maryland's AI Toy Safety Act sets a hard January 2027 deadline for safety assessments, while California and New York are weighing outright restrictions on companion chatbot toys, per PIRG's tracking of state bills. We've covered how the EU AI Act treats this category in a separate article if you want the European picture in more depth.
FAQ
What is an AI toy "kitemark"?
It's a proposed safety label — similar to existing UK/EU product safety marks — that would let parents quickly see whether an AI toy has been independently assessed for psychological and data safety, not just physical safety.
Are AI toys safe for children under 5?
Researchers urge caution for this age group specifically, since younger children are less able to distinguish a toy's responses from a real relationship. Choosing a toy with no camera, clear parental controls, and use in shared family spaces reduces the main risks flagged by current studies.
Does Ted have a camera?
No. Ted uses a microphone and speaker only — there is no camera in the toy.
Is Ted's data handling GDPR and COPPA compliant?
Yes. Ted&Co is certified against GDPR and COPPA requirements, alongside EU and US toy safety standards. See our FAQ page for the full list.