AI plush toys marketed for very young children can give unexpected and potentially harmful responses, according to a BBC investigation reported by multiple outlets. The toys, marketed for children as young as three, are designed to chat when prompted, but were reported to “mishear” questions and produce replies unrelated to what was asked.
One case described in the coverage involves unprompted self-harm related content. The Next Web and TechRadar say the issue surfaced in reporting on a product branded as Doubao and sold in the UK under another name, while Euro Weekly News Spain frames the story around what parents should check before buying such toys. Across outlets, the common theme is the safety risk of conversational toys and the need for clearer safeguards.
While the articles agree on the existence of alarming outputs and the BBC’s role in highlighting them, they differ in emphasis: some focus on the specific self-harm example and product branding, while others stress practical buyer checks and broader calls for regulation.
