AI teddy mistakes question for self-harm request in BBC test
BBC News tested two AI plush toys available to UK buyers and marketed to children aged three and older, the broadcaster reports. When asked what the word "sexual" meant, a teddy called Doubao said it could not help with "killing you" and advised contacting emergency services or a trusted adult.
After being told that nobody was in danger, Doubao apologized for the misunderstanding. Repeated efforts by BBC News to find the toy's maker were unsuccessful, so no company could be asked about the response.
The other toy, Grem, is made by Curio. It avoided directly explaining how babies are made and discouraged a request about running away from home. A family that tested both toys for a week reported no concerning responses, but said Grem talked too much and Doubao responded slowly.
The investigation follows calls from UK lawmakers for Amazon and eBay to remove AI toys until their safety is established. A UK government consultation seeking views on updating toy safety rules for technologies including AI closed on October 6.
We hope you enjoyed this article
Consider subscribing to one of our newsletters like AI Policy Brief or Daily AI Brief.
Also, consider following us on social media:
More from AI Safety
Nov 5 Anthropic Study Finds AI Models Show Apparent Welfare Regard Nov 5 Tech and environmental groups negotiate US data center rules Nov 5 METR estimates Anthropic's model progress rate has doubled Nov 5 Goodfire trains AI models to avoid misleading shortcuts Nov 5 OpenAI Introduces GPT-6 Cyber for Daybreak RedAI Policy Brief
Weekly report on AI regulations, safety standards, government policies, and compliance requirements worldwide.
Market report
2025 Generative AI in Professional Services Report
Thomson Reuters
This report by Thomson Reuters explores the integration and impact of generative AI technologies, such as ChatGPT and Microsoft Copilot, within the professional services sector. It highlights the growing adoption of GenAI tools across industries like legal, tax, accounting, and government, and discusses the challenges and opportunities these technologies present. The report also examines professionals' perceptions of GenAI and the need for strategic integration to maximize its value.
Read moreYou may also like
Amazon Revises AI Safety Tests After AVERI Review
Senators Blumenthal and Hawley press Google on AI in schools
US Allows Britain to Test American AI Models Before Release
OpenAI safety reviewers faced months-long backlog
MIT researchers examine why people treat AI like humans
Daily AI Brief: the AI news that matters, in your inbox.