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

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

https://nz.news.yahoo.com/child-safety-lab-launching-independent-090051059.html

Publish Date: 2026-05-05 05:00:00

Source Domain: nz.news.yahoo.com

Here is a summary of the article in an unordered list:

– Common Sense Media has launched the Youth AI Safety Institute, an independent research and testing lab focused on evaluating the risks AI tools pose to children and teens.
– The institute aims to set safety benchmarks for tech firms and provide parents and families with information about various AI tools.
– Unlike current third-party AI safety organizations, the institute focuses on consumer-friendly safety ratings rather than broader societal risks.
– The institute will conduct “red teaming” to stress test AI models and identify potential safety risks, with results published as consumer-friendly guides.
– Initial funding of $20 million annually comes from donors including tech companies like OpenAI, Anthropic and Pinterest.
– The institute will develop AI youth safety standards and benchmarks which tech companies can incorporate into their development for improved safety.
– Unlike physical products, AI models can frequently update, making frequent and robust research vital to keep up with progression.
– The institute aims to avoid repeating the pitfalls of the social media era, where years passed before risks to young people were fully uncovered.
– The institute hopes public pressure, its own research, and industry connections will encourage tech companies to adopt the developed benchmarks.