AI trains on kids’ photos even when parents use strict privacy settings

Australian kids need more AI protections

Australia's Privacy Act was introduced in 1988 and has undergone several updates as technologies evolved. But reforms Australia's attorney general, Mark Dreyfus, announced will be released in November, will likely be the most sweeping updates yet, enhancing privacy online and giving individuals direct legal remedies following data breaches.

After these changes, platforms will likely be required to disclose how Australians' personal information is used. And platforms will likely have to conduct assessments for some data collection in order to prevent riskier data breaches of particularly sensitive data, such as biometric or facial recognition data.

These "important" reforms "will help enhance individuals' control over their personal information," Dreyfus promised, confirming that "the speed of tech innovation and the rise of artificial intelligence underpins the need for legislative change."

But Dreyfus' speech said nothing else about AI and very little about children's online privacy, only noting that "87 percent of parents want more legislation that protects children's privacy." Han told Ars that it's important for officials to "figure out how to protect a child's full range of rights in the digital world." To Han, this means not just protecting kids' privacy, or "thinking about their right to access information," but "the entire scope of" children's rights online.

HRW has urged Australian lawmakers to build in more protections for children against AI. They've recommended that the Children’s Online Privacy Code should specifically prohibit "scraping children’s personal data into AI systems" and "nonconsensual digital replication or manipulation of children’s likenesses." And to ensure that tech companies are accountable, the policy should also "provide children who experience harm with mechanisms to seek meaningful justice and remedy." These protections should be extended to all people in Australia, HRW recommended, but "especially for children."

“Generative AI is still a nascent technology, and the associated harm that children are already experiencing is not inevitable,” Han said. “Protecting children’s data privacy now will help to shape the development of this technology into one that promotes, rather than violates, children’s rights.”