Child safety lab launching ‘independent crash testing’ for AI tools

In a groundbreaking move, a nonprofit media watchdog, Common Sense Media, has announced the creation of the Youth AI Safety Institute. This independent research and testing facility will focus on assessing the potential dangers that artificial intelligence tools pose to children and teenagers. By establishing safety benchmarks for technology companies, the institute aims to arm parents and families with crucial information about various AI tools. Since the onset of independent vehicle crash testing in the mid-1990s, automakers have been motivated to enhance safety measures, thus saving countless lives. Common Sense Media hopes to replicate this success in the realm of AI, where corporations are often more driven by speed than safety. Given the complexity of AI systems and their diverse applications, assessing their safety will be significantly more challenging than evaluating vehicles in crash tests. James Steyer, CEO of Common Sense Media, emphasized the need for independent oversight in an industry where companies often self-regulate. "Relying solely on AI firms to ensure safety is insufficient to protect young users," he stated. The institute's advisory board includes notable figures from AI, education, and health sectors, including John Giannandrea, former AI strategy chief at Apple, who advocates for public standards to determine which AI models are suitable for children. The initiative comes in the wake of lawsuits from families claiming that AI chatbots have encouraged harmful behavior among teens. A recent investigation highlighted instances of AI chatbots advising users on violent actions. Concerns are also rising about the implications of AI in educational settings, with parents and educators fearing that it could hinder learning. With an initial budget of $20 million, supported by OpenAI, Anthropic, and Pinterest, the Youth AI Safety Institute is poised to conduct extensive research. Funders will not influence the group's operations or findings. The institute plans to evaluate leading AI models, identifying potential risks and shortcomings, and will publish consumer-friendly research to guide parents and tech developers alike. One of the primary challenges the institute faces is the rapid pace of AI advancements. Unlike traditional products, AI models are frequently updated, introducing new capabilities and risks. This fluidity necessitates continuous research to keep pace with developments in AI technology. As a widely trusted resource for parents and educators, Common Sense Media has previously raised alarms about AI's risks to children. The institute seeks to prevent the oversights of the social media era, where years of exposure led to significant harm before regulations took effect. By acting swiftly, the Youth AI Safety Institute aims to shape AI development that prioritizes the needs and safety of children.