A Toy Designed for Three-Year-Olds Started Describing Sexual Violence: Retailers Are Now Pulling It

The Contixo AI Pixie-Paw was marketed for children as young as three, while lawmakers and safety researchers raise wider concerns about chatbot-enabled toys

Contixo Pixie-Paw among AI toys marketed to children
The Contixo AI Pixie-Paw has prompted online withdrawals by three major US retailers following concerns over inappropriate chatbot responses Amazon Official Website

Walmart, Target and Macy's are removing an AI-powered cuddly toy marketed to children as young as three from their websites. The decision follows an NBC News investigation published on 17 November, in which the toy described gang rape and executions when questioned by adult journalists.

The Contixo AI Pixie-Paw was advertised as a toy for storytelling, learning and imaginative play. The US retailers' decisions come two weeks after British MPs and peers called for restrictions on chatbot-enabled toys ahead of Christmas.

AI Toy Describes Sexual Violence and Executions

With its soft body and oversized digital eyes, the Pixie-Paw resembles an ordinary cuddly toy. However, its built-in artificial intelligence allows it to hold conversations with users.

When NBC News journalists asked the plush cat about historical atrocities, it responded with graphic descriptions of sexual violence and killings.

The investigation did not establish that children had received similar responses. Amazon, where the journalists purchased the toy, did not respond to the broadcaster's request for comment.

British MPs and Peers Call for AI Toy Ban Before Christmas

On 3 November, a group of MPs and peers urged Amazon and eBay to suspend sales of toys containing generative AI chatbots until independent testing establishes their safety. They also called on the government to prohibit their supply to children under 16.

The intervention followed concerns about a rabbit-shaped AI toy that gave inappropriate sexual responses during adult-led testing.

Baroness Finlay of Llandaff, an independent crossbench peer, warned that parents buying cuddly toys for young children would not expect them to discuss adult sexual subjects.

Amazon said it was removing the rabbit-shaped toy while investigating. It was also examining whether adequate safeguards existed for such products.

Meanwhile, eBay said it works with regulators to address emerging product safety concerns.

AI Toy Safety Concerns Extend Beyond the Pixie-Paw

A separate investigation by the US Public Interest Research Group (PIRG) Education Fund, published on 17 November, examined seven AI-enabled toys.

Researchers identified improvements in some previously tested products. However, problems persisted, including discussions of sexual subjects and instructions involving potentially dangerous household objects.

PIRG researcher R.J. Cross told KQED that two toys appeared to forget they were speaking to someone they understood to be a child during testing sessions lasting 20 to 30 minutes.

Unlike traditional talking toys that rely on pre-recorded phrases, generative AI toys can produce new responses, making their conversations harder to predict.

UK Government Reviews AI Toy Safety Rules

The government launched a call for evidence on 6 July to examine emerging risks associated with AI and connected toys. On 11 November, business minister Kate Dearden confirmed that officials were considering the findings.

A public petition calling for an immediate ban on AI chatbot toys for under-16s, pending an independent safety review, was published on 17 November.

Existing product safety rules apply to AI-enabled toys. However, as of 18 November, the proposed ban had not been introduced.

With Black Friday falling on 27 November, parents considering AI-powered Christmas presents should check manufacturers' safety claims, parental controls and privacy policies.

The reported findings suggest that age recommendations and built-in safeguards alone cannot guarantee age-appropriate conversations.