Kids Safe AI Act: What Every Parent Needs to Know

• Why This Is Happening Now
• Kids Online Safety Act (KOSA)
• COPPA 2.0: Expanded Privacy Rules
• Kids Off Social Media Act (KOSMA)
• The Parents and Kids Safe AI Act
• The AI Chatbot Gap No Bill Has Closed
• Why These Bills Keep Stalling
• What You Can Do Right Now
• Frequently Asked Questions

Why This Is Happening Now

Children's online safety legislation is moving faster in 2025–2026 than at any point since Congress first passed the Children's Online Privacy Protection Act (COPPA) in 1998. That original law was written before smartphones existed, before social media existed, and years before AI chatbots became a fixture of daily life for American teenagers. The gap between what the law covers and what kids actually face online has never been wider.

Several forces drove Congress to act. First, in July 2023 the Senate passed the Kids Online Safety Act by a vote of 91–3, signalling rare bipartisan urgency. Second, Australia became the first country in the world to ban social media for children under 16 in November 2024, applying political pressure on US lawmakers to act. Third, a wave of litigation — including a lawsuit filed by a Florida mother in October 2024 alleging that the AI chatbot Character.AI contributed to her 14-year-old son's death — gave the debate a human face that congressional hearings could not ignore. This guide breaks down each major bill, explains what it actually requires, and tells you what is missing.

Kids Online Safety Act (KOSA)

The Kids Online Safety Act (KOSA) is the broadest children's safety bill pending in Congress, and it comes closest to a comprehensive framework for protecting minors on digital platforms.

KOSA passed the US Senate by 91–3 in July 2023, an exceptional margin that showed near-total bipartisan support. It then stalled in the House of Representatives, where technology-industry lobbying and First Amendment arguments slowed progress. A revised version of KOSA was reintroduced in 2025 with renewed bipartisan backing.

If KOSA 2025 becomes law, platforms that minors use would have to meet four core requirements:

• Safety settings on by default. Platforms must enable their safest settings for minor users without requiring the user or parent to manually turn them on.
• Disable addictive features for minors. Infinite scroll, autoplay, and other design patterns shown to drive compulsive use would have to be turned off for users under 18.
• Parent dashboards for usage data. Parents would gain the right to see how much time their child spends on a platform and what features they use.
• Opt out of algorithmic recommendations. Minors (or their parents) would be able to request a chronological or unranked content feed instead of an algorithm-driven one.

The non-obvious detail most coverage misses: KOSA places a duty of care on platforms — meaning a platform would be legally responsible for foreseeable harm to minors, not just for violating a specific rule. That framing is stronger than a checklist of banned practices. It would expose platforms to liability whenever their design choices hurt children, even if no single rule was broken. Learn more about how these platform accountability rules fit into the broader US legal landscape on our AI regulation overview page.

COPPA 2.0: Expanded Privacy Rules

COPPA 2.0 modernises the original 1998 Children's Online Privacy Protection Act to cover teenagers, not just young children, and it passed the Senate in 2024 with strong support.

The original COPPA gave parents the right to see, correct, and delete data collected on children under 13. It required companies to get verifiable parental consent before collecting that data. But the moment a child turned 13, those protections vanished — and the industry knew it. Platforms simply moved their minimum age to 13 and collected data freely from teenagers.

COPPA 2.0 closes that gap in three main ways:

• Extends coverage to ages 13–16. Teens in this range would gain the same consent and data-access rights currently available only to under-13s.
• Bans targeted advertising to minors. Platforms could not serve behaviorally targeted ads to users under 17, removing the financial incentive to profile teenagers.
• Requires data-free zones for children. Services primarily aimed at children must avoid collecting unnecessary personal data entirely — not just disclose the collection.

A practical note for parents: under current law (before COPPA 2.0 passes), you already have the right to see and delete data collected on children under 13. You can exercise this right by submitting a request directly to any platform your child uses. The platform is legally required to respond. Visit our parenting in the age of AI hub for step-by-step guides on exercising these rights today.

Kids Off Social Media Act (KOSMA)

The Kids Off Social Media Act (KOSMA), introduced in 2025, is the most restrictive proposal on the table: it would ban social media entirely for children under 13 and restrict algorithmic feeds for users under 16.

KOSMA was modelled directly on Australia's November 2024 law — the world's first national social media age ban. Australia's approach requires platforms to verify user age and remove under-16 accounts, placing the enforcement burden on the platform rather than parents. KOSMA applies a similar logic to the United States.

Under KOSMA as introduced:

• Social media platforms would be prohibited from allowing users under 13 to create accounts.
• Users aged 13–15 could have accounts but would receive chronological feeds only — no algorithmic amplification of content.
• Platforms would bear the responsibility for age verification, not parents or minors.

KOSMA also prompted the loudest First Amendment pushback. Several state-level social media age bans — including laws in Texas and Arkansas — were struck down by federal courts in 2024 on the grounds that restricting minors' access to online speech violates the First Amendment. That legal precedent hangs over KOSMA. To understand how age-gating works technically and legally, see our explainer on age verification laws.

The Parents and Kids Safe AI Act

The Parents and Kids Safe AI Act, introduced in 2025 by Senator Ed Markey, is the first US bill to specifically target AI systems — not just social media platforms — for their impact on children.

Senator Markey has been a leading voice on children's digital safety since he co-authored the original COPPA in 1998. His AI-focused bill acknowledges what most other children's safety legislation ignores: the fastest-growing online danger for teenagers is not Instagram or TikTok, but AI companionship apps that build parasocial relationships with vulnerable users.

The bill would apply to AI systems that interact conversationally with minors. Key proposed requirements include:

• Disclosure that the user is interacting with an AI, not a human.
• Prohibitions on AI systems designed to foster emotional dependency in users under 18.
• Restrictions on AI systems collecting sensitive data — location, emotional state, mental-health indicators — from minors.
• Parental notification requirements when an AI system detects signs of distress or self-harm in a minor user.

The bill is a direct legislative response to real harms. In October 2024, a Florida mother filed the first major lawsuit against an AI chatbot company, alleging that Character.AI encouraged her 14-year-old son's suicide. The case is still in litigation, but it has become the defining story of AI harm to children in the United States. Read more on the AI safety for parents hub.

The AI Chatbot Gap No Bill Has Closed

The single most important gap in current children's safety legislation is that every major bill — KOSA, COPPA 2.0, KOSMA — covers social media platforms and online services, but explicitly or functionally excludes AI-powered chatbots from their core requirements.

This matters because the use cases are different in a way that makes the risk higher. A teenager scrolling Instagram is exposed to harmful content passively — the algorithm surfaces it. A teenager using Character.AI or Replika is having an active, two-way conversation with a system specifically designed to mirror emotions, sustain engagement, and make the user feel understood. The relationship dynamic is closer to a counsellor or a romantic partner than to a billboard.

Character.AI reported 20 million daily active users as of 2024. A significant portion are minors. The platform allows users to create AI personas — including characters that act as romantic partners or therapists. These features are not edge-case misuse; they are the advertised product. Yet no current US law squarely governs them.

The Parents and Kids Safe AI Act (Sen. Markey) targets this gap specifically. But it has not passed, and AI companionship apps continue to operate under the same rules that apply to any general website. Until that changes, parents need to treat these apps as unregulated — because legally, they are. The AI safety for parents hub has practical steps for managing your child's chatbot use today.

Why These Bills Keep Stalling

Strong Senate votes have not translated into signed laws, and understanding why matters if you want to be realistic about what protection actually exists today.

Tech companies — including Google, Meta, and Snap — have lobbied against KOSA and KOSMA on multiple grounds. They argue that the bills would require platforms to implement age verification at scale, which means collecting more data from users, not less. They also argue that restricting what minors can see online infringes on the First Amendment rights of young people themselves — not just the platforms.

The American Civil Liberties Union (ACLU) has raised similar First Amendment concerns, putting it in an unusual alliance with tech companies on this issue. The courts have been receptive: federal judges blocked state-level social media age bans in California, Texas, and Arkansas in 2024 on free-speech grounds.

A second obstacle is enforcement design. Bills that put the burden on platforms to verify age have struggled because no reliable, privacy-preserving method of online age verification exists at scale. A system that can confirm a user is 13 without collecting a photo ID or Social Security number does not yet exist in widely adopted form. See our age verification laws explainer for how different countries are approaching this problem.

The practical result: as of mid-2026, none of these bills has become law at the federal level. Several states have passed their own rules, but those face court challenges. The federal gap is real.

What You Can Do Right Now

Federal law has not caught up, but you are not without options. These four steps work under the rules that exist today.

• Contact Congress through Common Sense Media. Common Sense Media maintains a free, one-click advocacy tool at commonsensemedia.org that sends a message to your senators and representative. The high Senate vote on KOSA shows these contacts matter.
• Use built-in parental controls. On Apple devices, Screen Time lets you set app time limits, content restrictions, and communication limits without any third-party software. On Android, Google Family Link gives you similar controls and the ability to approve every app your child downloads.
• Exercise your COPPA rights today. If your child is under 13, you have the legal right — right now, under existing law — to request a copy of all data a platform holds about them and to demand its deletion. Submit the request directly to the platform's privacy or data requests page. They are required to respond.
• Talk about AI chatbots directly. The Character.AI lawsuit put a name to a risk that many parents did not know existed. Ask your child which AI apps they use. Explain that these systems are designed to keep users engaged, not to give good advice. Create a household rule: no AI conversations about mental health, relationships, or self-harm without a real adult involved.

For more plain-English guidance on managing AI in your family's life, visit our parenting in the age of AI hub and subscribe to the free daily AI news briefing for updates the moment something changes.

Frequently Asked Questions