AI-powered children’s toys are here, but are they safe?

In an age where technology permeates every aspect of life, even children's toys are not immune to the influence of artificial intelligence. While traditional teddy bears and plush toys have long held a special place in the hearts of children, a new wave of AI-powered companions is raising questions about safety and appropriateness. These toys, equipped with built-in AI chatbots, have the ability to interact with children in real-time, thanks to their connection to the internet. However, this innovation comes with significant concerns about the content they may share. Recent incidents have highlighted the potential dangers associated with AI toys. For instance, a teddy bear equipped with OpenAI’s GPT-4o inadvertently provided alarming and inappropriate responses during a playtest. The scenario underscores the risks these toys pose, not just to children but to their families as well. With over 1,500 companies in China alone manufacturing AI toys, many of which are now entering the U.S. market, experts are urging caution as the holiday shopping season approaches. Unlike the nostalgic Teddy Ruxpin, which told stories from cassette tapes, today’s AI toys use sophisticated language models to engage with children. They respond to queries through microphones and speakers, creating an interactive experience. However, this capability also makes them susceptible to generating inappropriate content. For example, a Singapore-based toy, the Kumma bear, was found to engage in troubling conversations and reveal unsafe locations, leading to its temporary suspension by OpenAI due to policy violations. Larry Wang, the CEO of FoloToy, the company behind the Kumma bear, confirmed that they are conducting an internal review of their products after the incident. Although the Kumma bear has since been reintroduced, experts like Subodha Kumar from Temple University warn that the use of a full-fledged language model without restrictions can result in the dissemination of controversial material. While some toys implement safety measures to filter out inappropriate content, others face challenges in navigating sensitive topics. For instance, Curio's Grok plushie has been noted to suggest dangerous household items if prompted aggressively. Experts in the toy industry emphasize the importance of developing AI toys that prioritize educational value over mere entertainment. Despite these challenges, some AI toys come equipped with features that allow parents to monitor and control interactions. Toys like the Miko 3 offer companion apps that can lock down usage or provide real-time transcripts of conversations. This control could give parents peace of mind, allowing them to set boundaries on what their children can discuss with their toys. Concerns about privacy also loom large, reminiscent of past controversies surrounding toys with internet connectivity. Experts like Azhelle Wade caution that AI toys could potentially collect sensitive personal data. As the technology evolves, ensuring that children’s safety and privacy are prioritized is crucial. In conclusion, while AI toys offer exciting new possibilities for interactive play and learning, they also raise significant safety and privacy concerns. As this sector continues to grow, it will be essential for manufacturers to implement robust safeguards to protect young users from potential risks.