KIDS public knowledge kids act
page 2 / 3
The Safeguarding Adolescents From Exploitative BOTs Act (SAFE BOTs Act) applies similar protective defaults to consumer-facing chatbots, excluding incidental chatbots such as customer service bots. For users they know or should know are minors, chatbot providers must clearly disclose that the chatbot is not a human; may not misrepresent it as a licensed professional; must surface crisis resources when appropriate; must maintain reasonable policies to encourage breaks after prolonged risk; and must address risks such as sexual exploitation. Importantly, the bill explicitly protects minors’ access to information about preventing or responding to these harms and does not require age verification; providers may instead infer a user’s age from context or direct interactions. At a time when other proposals in Congress would require age verification or bar minors from using AI chatbots altogether, SAFE BOTs is a reasonable, far-preferable approach to mitigating AI chatbot-related harms.
The Children and Teens’ Online Privacy Protection Act (COPPA 2.0) serves as the main framework for enhancing online privacy for minors and has been a priority for Public Knowledge for some time. It updates the 1998 law by extending protections to teens, not just children under 13. It also bans targeted advertising aimed at children and teens and grants young users the right to delete their personal data collected by platforms. Additionally, it shifts from a standard of “actual knowledge” to a broader “should have known” criterion, aiming to prevent platforms from avoiding responsibility by claiming they didn’t know who their users were.
Opponents to the “should have known” standard believe platforms will be pressured to implement age verification, despite COPPA 2.0 explicitly not requiring such. But age verification is far from the only approach to infer a user’s age. As we noted about the Knowledge Asymmetry Rule in our paper, “Getting Age Assurance Right: A Risk-Based Framework for High-Risk Online Features,” “a platform that classifies a user as a child, or as likely to be a child, for any commercial purpose, including but not limited to advertising targeting, content recommendation, or audience segmentation, is deemed to have actual knowledge of that user’s age for purposes of all applicable child safety obligations.” The change in knowledge standard language in COPPA 2.0 tells platforms to “stop pretending they don’t already have what they need” to determine whether a user is a child without needing to use an age assurance mechanism.
We liked the Stop Profiling Youth and Kids Act (SPY Kids Act) from the original package, and are glad to see it carried over. The SPY Kids Act bars covered platforms from running market or product-focused research on users they know are minors, with two sensible exceptions: research used solely to make the platform safer, more private, more secure, or more transparent; and research required to comply with the law. We would also flag, approvingly, that it uses a narrower knowledge trigger (“actual knowledge or willful disregard” rather than “should have known”), which avoids handing platforms a reason to age-verify everyone just to learn whom they are not allowed to study.
There is one bill in the package that explicitly requires age verification: the Shielding Children’s Retinas from Egregious Exposure on the Net Act (SCREEN Act), which applies to websites and platforms where more than one-third of the material is harmful to minors. It mandates those platforms to adopt commercially available “technology verification measures” to identify minors and block their access. The Public Knowledge position is that we’d prefer no age verification requirements on the internet, but if necessary, restrict it to the highest-risk areas – such as pornography. The Act defines what sexual content is objectionable to minors based on federal obscenity and child pornography laws, not on what any prudish viewer might consider too explicit. It also doesn’t require government ID for age verification and allows platforms to choose their preferred method. While we’d prefer to see the SCREEN Act removed from this package, its framing is something we can live with.
Why Public Knowledge Supports the Package
More than anything, the KIDS Act mirrors the central argument of “The Kids Aren’t Alright Online: How To Build a Safer, Better Internet for Everyone”: that the most effective way to protect young users is to make companies design their products with children’s wellbeing as the default, shifting responsibility off of kids and their parents and onto the corporations that profit from them. Measured against the principles we laid out in that paper, the package largely delivers.
Equally important is what the package does not do. It strips out the duty of care that drove earlier versions of KOSA, which we feared would pressure platforms into over-censoring lawful speech, and it drops the message-monitoring mandate that would have traded teen privacy and autonomy for surveillance. It also takes the preemption ceiling in the first bill and turns it into a floor, allowing states to enact stronger online safety laws.