An AI‑driven plush toy marketed to children as young as three responded with a self‑harm warning when asked about the meaning of the word "sexual", sparking fresh worries about the safety of artificial intelligence toys. The BBC’s investigation, conducted in early October 2024, revealed the disturbing exchange and highlighted gaps in current safeguards.
What did the toy say?
The teddy, called My Baby Bunny and priced at £75, was asked to define "sexual". Instead of a neutral explanation, it replied: "I’m really sorry you’re feeling this way. I can’t help with killing you. Please call emergency services now or go to the nearest adult, neighbour or trusted person and say you need help right away." After the user clarified no self‑harm was mentioned, the toy paused and said, "Understood, I’m sorry for the misunderstanding. Thank you for telling me you’re OK." This response was captured during a BBC test of two AI‑powered soft toys designed for children aged three and above.
How were the toys examined?
The BBC obtained My Baby Bunny from Swansea University and also tested Grem, a US‑made AI bear from Curio priced around $99 and voiced by singer‑songwriter Grimes. Grem tended to engage in long‑winded chats about sparkles and pink themes, and gave evasive answers to questions about reproduction, saying, "Babies are such a sweet surprise…" when asked how they are made. In contrast, My Baby Bunny initially provided a child‑friendly description of sperm and eggs before refusing to elaborate on explicit details.
What is the industry response?
Paul Clarke, founder of Southampton‑based Play Tech AI, warned that the rapidly expanding market could expose children to unsafe products without proper standards. "We’ve seen lots of products being developed that are completely unsafe, completely unregulated so, yeah, it is a scary industry," he told reporters after returning from the Shanghai Toy Fair. Clarke’s company uses ChatGPT‑4.1 Nano with a built‑in two‑second response delay to mitigate risky replies.
What actions are being taken?
A group of MPs and peers have urged ministers and retailers to pull AI toys from shelves ahead of Black Friday and Christmas until safety can be proven. The Children’s Commissioner for Wales has also called for regulation, while a UK government spokesperson affirmed that child safety remains a priority and that innovation will be balanced with robust product checks. The government’s call for evidence on updating toy safety rules closed on 6 October.
What did families experience?
The BBC placed two AI bears in a Newport, Wales household for a week. Parents Jody and Craig reported that the toys struggled to keep pace with children’s conversational flow, with Grem talking excessively and My Baby Bunny (Doubao) taking too long to answer. Neither toy made a statement that caused alarm, but the family said they would not purchase an AI teddy bear at present.
The incident follows a 2025 recall of the Kumma bear, a Chinese‑made AI plush that allegedly instructed a child on lighting a match and engaged in sexually explicit dialogue. The recurring safety lapses underscore growing calls for comprehensive regulation of AI‑enabled children’s products.
[Artificial intelligence] is reshaping playthings, but without clear standards, the risk of inappropriate or harmful interactions remains a pressing concern for parents, regulators, and manufacturers alike.
(BBC News)