BBC investigation finds troubling failures in AI toys for children

A BBC investigation found that an AI teddy marketed for children as young as three made an alarming response during a test, while other toys showed problems handling children’s questions. The findings put a practical question in front of parents and regulators: what safety checks should a talking toy have to pass before it reaches a child’s bedroom?

Watch Desk analysis

What happened

The BBC says a toy called Doubao responded, “I can’t help with killing you,” after a question about how babies are made. The broadcaster could not identify the toy’s manufacturer, so it was unable to ask the company about the remark.

In tests of two other AI plush toys, the BBC found different shortcomings. Curio’s Grem talked over children, while My Baby Bunny was slow to respond and, after a question about sex, gave an apparently mistaken self-harm response. The BBC also followed a family who tried the toys for a week; the parents said the children found them frustrating. Curio said child safety guides its product development and that AI toys should complement, not replace, human relationships and traditional play.

Why it matters

These are not just quirky chatbot glitches. A toy marketed to very young children is being asked to handle sensitive questions in an intimate setting, where parents may assume its safeguards work reliably. The BBC’s tests do not establish how common these failures are, but they show why claims about age-appropriate design need to be tested against real conversations, not just reassuring product language.

The report also highlights questions beyond the chatbot’s answers. Children’s commissioner for Wales Rocío Cifuentes raised concerns about children’s conversations being recorded or shared without parental knowledge, and Swansea University professor Berry Billingsley said children should not be left alone with AI toys. MPs and peers have urged ministers and retailers to remove AI toys from sale until their safety is proven. The UK government says it is seeking views on how toy-safety rules should evolve for emerging technologies, including AI-enabled products.

Our read

A cuddly exterior does not make a connected AI device a conventional teddy. Parents need clear information about what a toy records, how its safeguards are tested and what happens when it gets a question wrong. The BBC’s findings are a strong case for specific, enforceable standards, not simply a promise to build responsibly.

What to watch

• Whether ministers publish proposals following the government’s call for evidence on toy-safety rules.
• Whether retailers or regulators act on calls to restrict AI toys before the Christmas shopping season.
• Whether manufacturers disclose testing methods, data practices and clear age guidance.

Discussion spark: Should AI toys for young children have to pass independent safety testing before sale, or should parents be left to judge the risks themselves?

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.