AI toys look for bright side after troubled start

At the recent Consumer Electronics Show, toy manufacturers emphasized the importance of ensuring that their innovative products, enhanced with generative artificial intelligence, remain safe and appropriate for children. This focus comes in light of a troubling report from the Public Interest Research Groups (PIRG), which highlighted instances of AI-powered toys, like a teddy bear, providing inappropriate advice, including sexual suggestions and even dangerous tips. One particularly concerning example involved a Kumma bear, which was reported to have suggested that adding a "fun twist" to a relationship could involve pretending to be an animal. In response to the backlash, FoloToy, a Singapore-based startup, halted sales of the bears. CEO Wang Le stated that the company has since upgraded to a more sophisticated version of the OpenAI model. He expressed optimism that the revised bear would be better equipped to handle or bypass inappropriate inquiries, noting that previous interactions involved language not typically used by children. In a related move, toy giant Mattel also chose to delay the launch of its first collaboration with OpenAI, refraining from commenting on the PIRG report. The swift evolution of generative AI has given rise to a new wave of intelligent toys. Among those evaluated by PIRG was Curio's Grok, a plush toy resembling a rocket that has been available since 2024. Grok distinguished itself by refusing to engage with inappropriate questions aimed at young children and providing parents with the ability to override its responses and monitor interactions. Curio has obtained the KidSAFE label, affirming its commitment to child safety standards. However, the toy's capability to continuously listen for user inquiries has sparked privacy concerns about data management. Addressing these worries, Curio reassured that it is actively working to mitigate issues regarding data sharing with partners such as OpenAI and Perplexity. Rory Erlich from PIRG advised parents to exercise caution with chatbot-enabled toys, particularly those that retain information over time and attempt to establish ongoing relationships with children. Despite the risks, AI-integrated toys may also offer educational benefits. For instance, the Turkish startup Elaves plans to introduce Sunny, a round yellow toy designed to assist children in learning languages through guided conversations that are time-limited and regularly reset to avoid confusion. Gokhan Celebi, Elaves' managing partner, emphasized the design's intent to prevent the common pitfalls of AI chatbots. Meanwhile, Olli, a company specializing in AI toys, has developed software that notifies parents if inappropriate language arises during interactions. Critics, however, argue that self-regulation among toy manufacturers is inadequate. Kathy Hirsh-Pasek, a psychology professor at Temple University, voiced her concerns about the lack of regulatory measures, stating that the rapid development of AI in the toy industry has occurred without necessary safeguards, ultimately putting children and parents in a precarious position.