US Senate Approves New Bills Targeting Kids' AI Safety, Creating New Compliance Deadlines

The US Senate Commerce Committee recently approved the Kids Online Safety Act and several other bills focused on children's safety, shifting the regulatory landscape for AI products used by young people.

What Happened

The committee advanced the Kids Online Safety Act, known as KOSA, along with companion bills. These measures would require AI platforms and online services to actively design their systems to protect children from harm. This includes risks like cyberbullying and the exposure to inappropriate content. The bills focus on the design phase of the product. They mandate that companies prioritize safety by default when building AI tools for minors.

The Risk Failure Mode

The failure here is the reactive nature of current risk management. Companies often add safety features after users complain or after regulators issue fines. These bills force a proactive approach. The risk now is non-compliance with new design obligations, which could lead to significant fines or bans. This is similar to the Clearview AI case, where mass facial-image scraping drew multiple GDPR fines. The new bills signal that regulators expect companies to build safety into the AI from the start, not patch it in later.

Regulatory Landscape

Unlike the EU AI Act, which is a horizontal law covering all industries, these US bills are sector-specific. They target social media platforms and online services. The mechanism is a duty of care that extends to the design choices of the AI system. Companies must be able to demonstrate that they have assessed the risks to children and taken steps to mitigate them. This creates a new category of compliance work for risk teams.

Key Controls

Companies need to map their AI products to the specific requirements of these new bills. They need to document their design decisions to show they prioritized child safety. They need to implement specific safety features like default privacy settings for minors. This requires a shift in how AI is built, moving from a focus on engagement to a focus on protection.

What to do

• Map your AI products to the specific requirements of KOSA and related bills
• Conduct a fundamental rights impact assessment focused on minors
• Update your design process to prioritize safety by default
• Prepare for a patchwork of state-level rules if federal rules are delayed
• Document your safety design decisions for regulatory review

More from our platforms

These sister platforms cover the parts of this problem that sit outside governance.

• Argus (argus.threatclaw.ai) records every trace an AI application produces and scans it for prompt injection, jailbreaks and data leaks, including the attacks hidden inside retrieved documents and tool results rather than in what the user typed. Governance decides what an AI agent is allowed to do. Argus shows what it actually did.
• ThreatClaw (www.threatclaw.ai) tracks the threat side of the same systems: 22 live intelligence feeds, exploitation predicted before it is officially confirmed, threat actor profiles, and detection rules you can deploy straight away. A control is only as good as the threat it is sized against.