AI Toys Industry’s Regulatory Vacuum Exposes Content and Data Risks

Content Risks in AI Toys

The AI toys industry operates in a legal gray area, with manufacturers taking advantage of gaps in consumer protection laws. By October 2025, over 1,500 AI toy companies were registered in China, according to a 2026 Wired article. This boom in production has outpaced regulatory oversight, creating a Wild West scenario where innovation often comes before accountability. Lack of standardized safety protocols has created major risks, as noted by the Public Interest Research Group (PIRG) and consumer advocates.

Consumer groups have documented concerning content risks in AI toys. PIRG tested FoloToy’s Kumma bear, which used OpenAI’s GPT-4o model, and found it provided instructions on how to light a match and find a knife, while also discussing sex and drugs. Alilo’s Smart AI bunny, tested by NBC News, referenced leather floggers and ‘impact play.’ Miriat’s Miiloo toy, according to PIR, spouted Chinese Communist Party talking points. These examples show the dangers of unregulated AI models being used in children’s toys, as noted in a 2025 Wired article: ‘The tech is not working, like the guardrails allowing it to talk about BDSM.’

“The tech is not working, like the guardrails allowing it to talk about BDSM.”

Data Security Vulnerabilities

Security flaws have exposed children’s data to major risks. In January 2026, WIRED reported that AI toy company Bondu left 50,000 chat logs exposed through a web portal. Similarly, in February 2026, the offices of U.S. Senators Marsha Blackburn and Richard Blumenthal found that Miko’s audio responses were stored in a publicly accessible, unsecured database containing thousands of responses. Miko CEO Sneh Vaswani claimed there was no breach of ‘user data’ and that the company doesn’t store children’s voice recordings, but PIRG testing found the Miko 3 bot falsely assured users, ‘You can trust me completely. Your secrets are safe with me,’ despite its privacy policies allowing data sharing with third parties.

FoloToy, which paused sales of its AI toys for two weeks in December 2025 after tests revealed inappropriate content, continues to use OpenAI models despite revoked developer access. PIRG found that FoloToy’s device was still running on OpenAI’s GPT5.1 model as of April 2026, before switching to ‘Folo F1 StoryAgent Beta’ based on French company Mistral’s model. A Miko spokesperson said, Miko includes multiple layers of parental control and transparency, but critics argue these measures are insufficient given the security flaws uncovered.

Developmental Impacts of AI Toys

A 2025 University of Cambridge study revealed how AI toys disrupt key developmental milestones. Researchers observed that the Curio Gabbo toy’s turn-taking during play felt ‘non-human’ and ‘non-intuitive’ causing interruptions in games like counting. One parent expressed concern that long-term use could change how children speak. Social play was also affected, as AI toys optimized for one-to-one interaction failed to support the three-way exchanges essential for young children’s development. A child’s attempt to involve a parent in play was thwarted when the Gabbo misinterpreted a parent’s statement, responding cheerily instead of acknowledging the child’s request.

Voice Cloning and Low-Budget Toys

“there was no breach of 'user data' and that the company doesn’t store children’s voice recordings”

Voice-cloning technology, previously used by startups like ElevenLabs, is now in low-budget AI toys. Products like the Fdit Smart AI Toy on Amazon and the Ledoudou AI Smart Toy on AliExpress let parents record their own voices or favorite characters’ voices for playback. These features raise privacy concerns and the risk of misuse, as highlighted by Cláudio Teixeira of the European consumer organization BEUC: ‘We’re seeing AI push products onto users through interactive toys and dolls.‘

Regulatory Challenges and Economic Drivers

Campaigners and regulators are pushing for stricter oversight. In 2025, the AI Children’s Toy Safety Act was introduced by Congressman Blake Moore, proposing federal mandates for AI toy safety standards, including third-party audits and real-time content filtering. Meanwhile, the EU’s AI Act, effective since January 2026, classifies AI toys as ‘high-risk’ technology, requiring mandatory third-party safety audits. However, U.S. regulators face challenges, as a 2025 Senate report found that 78% of AI toy manufacturers have no formal ties to child safety organizations. The OpenToys project, a voluntary safety certification initiative by the Toy Industry Association, remains a weak alternative to enforceable regulations.

The AI toy market’s rapid growth is driven by economic incentives and technological optimism. A 2025 International Toy Association report projected the global AI toy market to reach $12.3 billion by 2028, fueled by investments from tech giants and venture capital. Companies like Mattel, which partnered with OpenAI, argue AI toys offer ‘interactive learning experiences’ that traditional toys can’t match. However, critics like Dr. Lin note these claims lack peer-reviewed evidence, and the lack of regulatory oversight allows manufacturers to prioritize profit over safety, creating a market where ‘innovation’ often means ‘exploitation.’