I work hard to be anti-alarmist, to find the middle path in most things. The place where I can be both pro-technology and pro-online safety. But this week is a time to sound an alarm.
Kids urgently need your help now.
Last year, the Senate passed the Kids Online Safety Act (KOSA) 91-3. The version that was just introduced in the House does not include the same protections that the Senate passed and is extremely problematic for two main reasons:
It removes the “duty of care” language which requires social media companies to design and maintain their products with kids’ safety in mind. This is not an issue of content or censorship, but a mechanism to ensure Big Tech can’t build products that optimize for profit or engagement when they understand real or potential harms to kids.
It includes broader preemption language on both kids online safety and AI legislation, aiming to undo steps taken and thwart future progress by states who have introduced and passed bills and Executive Orders that protect kids.
I’ll include a lot of additional context below, but TL;DR here’s what kids need from you:
Please contact your congresspeople and insist that:
The House uses the Senate version KOSA which includes a Duty of Care.
Representatives remove all preemption language related to AI legislation and kids online safety legislation.
Demand a federal floor, not a ceiling: When discussing any federal legislation on these topics, insist that Congress set a strong minimum standard but allow states to go further if needed.
You can call, email, or write a letter. I’ve heard directly from federal legislators that even 3 phone calls or 10 emails is enough to take constituent concern seriously. I’ve included a script for phone or email below.
Script for Contacting Congress:
Find the contact information for your congresspeople here.
Hello, my name is [Your Name], and I’m a constituent from [Your City, State]. I’m very concerned about children’s safety online. I’m calling to urge [Senator/Representative Name] to support stronger online safety measures and reject any attempts to weaken protections for kids.
First, please restore the “duty of care” provision in the Kids Online Safety Act (KOSA). The original Senate-passed KOSA included a duty of care requiring social media platforms to exercise reasonable care to protect children from harm. Unfortunately, the House version removed this crucial provision. Without a duty of care, tech companies won’t be fully accountable for reducing the harms their services can cause to minors.
Second, I ask you to oppose any blanket federal preemption of state online safety or AI laws, whether in KOSA, the National Defense Authorization Act (NDAA), or any other bill. We should not override the important protections that states have already put in place. Any federal law must set a strong floor of protections – a baseline that all must follow – while allowing states to go further. Please ensure that federal legislation only preempts state laws if they conflict, and explicitly allows states to enact stronger safeguards. Every single state introduced AI legislation in 2025, and 38 states enacted measures to protect consumers and kids. We need federal standards and the ability for states to respond to new threats.
Finally, I want to emphasize that protecting children online must come before Big Tech’s preferences. Too often, the tech industry pushes for weaker, one-size-fits-all laws or broad preemption to avoid stricter state rules. But our kids’ wellbeing is more important than any company’s convenience or profits. I was troubled to hear that some platforms “breathed a sigh of relief” when KOSA’s duty of care was removed. Please stand up to industry pressure and lobbying to put kids’ safety first.
(Optional: if you have a personal story or local example, add it briefly here to make your point. For example: “As a parent of two teens, I’ve seen how unchecked social media algorithms can push dangerous content. One of my children became withdrawn and anxious after encountering online bullying. This is why these protections are so important to me.”)
Thank you for your time and attention. I hope I can count on [Senator/Rep Name] to fight for strong online safety measures that protect our children and empower (not undermine) state efforts.
Kristin with a photo of her son, Carson. Source: NPR
Key Context: Duty of Care Provision
In the recent House proposal, the central “duty of care” provision was removed from KOSA – significantly diluting the bill’s mandate that platforms actively protect minors from harm. The original duty of care (included in the Senate-passed KOSA) required social media and other online services to mitigate a broad range of harms – from cyberbullying and sexual exploitation to eating disorders and self-harm content – or face liability for failing to shield underage users.
Parents who lost children to online harms championed this duty as a way to force platforms into safer design choices, but major tech firms and some lawmakers warned it could lead to overzealous content removal and free speech infringements, despite any content-related language being removed before the 2024 Senate voted version. Yielding to these already-resolved concerns, and to Big Tech lobbyists, House drafters replaced the duty of care with narrower requirements.
How we got here:
The Senate version of KOSA: Imposed an affirmative obligation on platforms to “protect kids from a variety of online harms” and act in children’s best interests. This meant companies could be held legally responsible if minors were exposed to content causing mental health issues (like eating disorders or suicidal ideation) or other serious harms due to the platform’s algorithmic or predatory design. Proponents – including bereaved parents and social media reform advocates like me – believe such liability would compel Big Tech to change algorithms and policies to prevent tragedies. Opponents – digital rights groups and some lawmakers, all supported by Big Tech’s own lobbyists – expressed fear that it would incentivize platforms to over-censor content, potentially sweeping up helpful materials (e.g. LGBTQ+ support resources) in an effort to avoid any risk. Again, content moderation is not a goal of KOSA, and LGBTQ+ kids are often most vulnerable to platforms’ product design negligence.
Who Pushed for Removal: Resistance to the duty of care came from House republican leadership and industry allies concerned about unintended censorship and lawsuits’ impact on business. After KOSA sailed through the Senate (91–3) in 2024, House Speaker Mike Johnson and Majority Leader Steve Scalise refused to bring it to a vote, calling the duty of care “very problematic.” After blocking KOSA in the House, Mike Johnson thanked Meta for their multi-billion-dollar data center investment in his home state of Louisiana. Representative Kathy Castor (KOSA’s Democratic co-sponsor) publicly criticized House leaders for “listening to Big Tech executives” over families’ concerns, suggesting industry lobbying influenced the overhaul.
Nature of the Changes: The House Energy & Commerce Committee’s new draft guts the duty of care, scrapping the expansive “protect against all harms” mandate. Instead, platforms must implement “reasonable policies, practices, and procedures” to address four specific categories of harm: (1) physical threats of violence, (2) child sexual exploitation and abuse, (3) illicit drug sales, and (4) financial fraud targeting minors. This pivot narrows the scope to clear-cut illegal or dangerous activities, omitting any direct requirement to tackle content linked to depression, eating disorders, or other mental health harms. It also depends on Big Tech’s definition of “reasonable,” which they have repeatedly demonstrated unable to adequately assess.
Reactions from Stakeholders:
Child Safety Advocates: Parent advocacy groups condemned the House’s move as a betrayal of the bill’s purpose. ParentsSOS, a coalition of families who lost children to online harms, blasted the new draft for dropping the “most crucial component” (the duty of care) that they had urged in every meeting. They warn that a “watered down” KOSA will fall short of preventing the kind of tragic outcomes they experienced. These advocates also object to a new preemption clause in the House bill (discussed below) that would block state online safety laws, calling it “attempting to stop states from protecting vulnerable children.”
“After three years of meeting with legislators to share our families’ stories of loss and our legislative priorities, we are extremely disappointed that this bill doesn’t contain the ‘duty of care’ we asked for in every meeting and that the bill preempts the state laws that so many parents have worked tirelessly to pass.”
Lawmakers: KOSA’s original bipartisan sponsors in the Senate, Sen. Richard Blumenthal (D-CT) and Sen. Marsha Blackburn (R-TN), insist the bill “needs” the duty of care and say there is “no appetite for watering the bill down.” Blackburn called establishing a duty of care “essential to protecting our kids and giving parents peace of mind.” In the House, co-sponsor Rep. Kathy Castor (D-FL) has echoed those sentiments, publicly urging GOP leadership not to deliver a watered-down KOSA and to spend “less time meeting with Big Tech executives.”
Big Tech: Major tech companies are “breathing a sigh of relief” at the House’s rewrite. The duty of care would have made platforms legally liable for teens’ mental health outcomes – a frightening prospect for companies like Meta and Google, which reportedly were “scrambling” to imagine how they’d avoid lawsuits over adolescent depression or eating disorders linked to their services. Removing this broad liability shielded them from what they viewed as an onslaught of unpredictable litigation.
I’ve worked inside the companies fighting these protections and written extensively about the failure of self regulation. Trust me: without a duty of care, they will not change their design choices.
Key Context: Federal Push to Preempt State AI and Kids’ Online Safety Laws
As the watered-down version of KOSA potentially advances, a parallel battle is unfolding over federal preemption of tech regulations.
In late 2025, some in Congress (and the White House) are attempting to block states from enforcing their own AI and online safety laws, arguing that emerging technologies require a unified national approach. This includes proposals for a moratorium on any new state laws regulating AI, as well as provisions in federal legislation (like the House KOSA draft) to invalidate existing state child-protection laws.
These efforts are being driven largely by the tech industry and its allies, who warn of a “patchwork” of 50 different rules hindering innovation. However, the pushback from states, bipartisan officials, and advocates argues that stripping state authority would undermine crucial protections for consumers and kids.
Parent advocates staged a protest on 12/1, projecting messages like “DON’T LET AI BUY THE GOVERNMENT” and “AI PREEMPTION HURTS KIDS” on the East Wing of the National Gallery of Art in D.C.. The protest was supported by online child safety advocates from Heat Initiative, ParentsTogether Action, and Young People’s Alliance. Source: Heat Initiative
Below is an overview of the key proposals and the debate:
10-Year Moratorium on State AI Laws (Failed Attempt): Over the summer of 2025, a sweeping proposal to ban states from regulating AI for a decade nearly became law. Senator Ted Cruz (R-TX) slipped a state AI moratorium into a must-pass budget package (nicknamed the “Big Beautiful Bill”). This provision would have blocked most state or local laws on AI outright for ten years.The rationale was to prevent “burdensome” local rules from stifling AI investment – a view enthusiastically backed by large tech firms and the Trump administration. However, the effort collapsed in the Senate: in July 2025, Senators voted 99–1 to strip the moratorium from the bill. Lawmakers from both parties raised alarms about its breadth and drafting flaws. Sen. Marsha Blackburn (R-TN) opposed the moratorium, warning that “blocking states [now] could allow Big Tech to continue to exploit kids, creators, and conservatives” in the absence of strong federal safeguards. Essentially, senators concluded that it was premature to forbid state action when Congress itself had not yet passed comprehensive AI or online safety rules. With near-unanimous agreement, the moratorium was removed from the budget bill.
Renewed Preemption Push via NDAA and Executive Action: The defeat of the moratorium has not ended the issue. By November 2025, House republican leaders were eyeing the National Defense Authorization Act (NDAA) – a must-pass defense bill – as a vehicle to “preempt state AI laws,” effectively reviving the ban through other means. Simultaneously, a leaked draft Executive Order from the White House proposed using federal funding leverage: it would withhold certain funds from states with “onerous” AI regulations deemed to “undermine [the] innovative culture.” The draft EO suggested tying state AI policy to broadband grants, penalizing states that don’t fall in line.
Notably, 300+ bipartisan state legislators from across the country sent a joint letter to congress this week opposing any revival of the moratorium, saying they routinely hear from constituents about AI harms and that “removing states’ ability to regulate… would be dangerous.”“A blanket prohibition on state and local AI laws would abruptly cut off active democratic debate… at the very moment communities are seeking responsive solutions,” those state lawmakers warned.
Similarly, a bipartisan coalition of 36 state Attorneys General sent a letter to congress this week that argued “broad federal preemption would undermine states’ ability to respond quickly and effectively to emerging AI risks.”
“The Senate killed a bill to ban state regulation of AI by a vote of 99-1 this spring. Now the industry - eager to stay unregulated - and Trump are scheming to shove the provision in the national defense bill. You need to know this, so you can help stop it.”
- Senator Chris Murphy
Preemption of State Kids Online Safety Laws (KOSA clause): The issue of state vs. federal authority isn’t limited to AI. Buried in the House’s new KOSA draft is a sweeping preemption clause that would “preempt states from enacting or enforcing their own laws to protect children” online. In plain terms, if the House version of KOSA passes, it could nullify dozens of state-level child safety initiatives. This includes recently passed “Kids Code” laws in states like Maryland and Nebraska, the Texas SCOPE Act (which restricts social media for minors), the New York SAFE Act, and various other laws nationwide that give parents and minors new digital rights.
It would also prevent any new state laws addressing social media harms or addictive design from being enacted, like those I’m deeply advocating for in my home state of Washington. Industry advocates and Big Tech applaud preemption here, noting that states have begun passing disparate online child safety laws that create compliance nightmares – a federal law could “solve [this] patchwork problem” (as they phrase it) and set a consistent bar across all 50 states.
“It’s bad enough that Congress hasn’t passed a law to protect children online in over 25 years. Now, they are attempting to stop states from protecting vulnerable children from online harms.”
- Sharon Winkler, mother of Alex Peiser, forever 17, and ParentsSOS founding member
Sharon with a photo of her son, Alex. Source: Fox5Atlanta
The alternative to these “compliance nightmares,” of course, is for Big Tech to hold themselves to the highest standard across the board. Wouldn’t it be great if state laws became unnecessary because Big Tech was actually creating safer products?
If Big Tech cares so much about avoiding “patchwork” legislation, they can raise their safety standards. Simple.
“While Big Tech lobbies for AI amnesty that would prevent states’ from , children across America remain vulnerable to the harms caused by AI, social media, and algorithmic exploitation. Congress should be focused on passing meaningful legislation that actually keeps kids safe online, not rewarding the industry’s inaction by stripping away state-level protections that protect our most vulnerable. States have stepped up to fill the void left by federal paralysis, and they should be applauded for protecting their children, not punished with preemption measures that serve corporate interests and tech billionaires over child safety.”
There are two main approaches under discussion: a time-limited moratorium (pause) on state laws, or permanent federal preemption through legislation. A moratorium would temporarily bar states from enacting or enforcing AI regulations for a set period (as Cruz’s 10-year ban attempted). Preemption, by contrast, writes into federal law that it “supersedes” state laws on a given topic, invalidating any state rules that conflict with the federal standard.
The current House proposals flirt with both: the NDAA idea appears to be a direct preemption (using the Supremacy Clause to nullify state AI laws), whereas the earlier failed BBB amendment functioned as a moratorium tied to funding. Both approaches seek to achieve the same end – preventing the states from regulating tech – but they operate differently.
Bottom line:
The decision Congress makes on preemption will have sweeping implications for the balance of power between federal and state regulators in tech. If a moratorium or broad preemption is enacted without accompanying strong federal rules, it could effectively freeze regulatory progress in areas like AI ethics and kids’ online safety.
States have long been the “laboratories of democracy” – pioneering new protections (e.g. the first privacy laws, biometric safeguards, teen online safety laws) which sometimes spur federal action. Removing that dynamism might please industry, but it eliminates the most responsive level of government for emerging digital harms.
Congress must act boldly to protect children online – but not by undercutting those protections at the state level. By raising these issues with your legislators, you help ensure that any final policy truly puts kids’ safety first, maintains vital local initiatives, and holds tech companies to account.
Lawmakers should hear that we want both: strong federal safeguards and the freedom for our states to innovate further.
We can’t put Big Tech’s compliance convenience over the safety and wellbeing of kids.
Overturned by Kelly Stonelake is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
This post is very well-written and is an accurate summary of the current state. Thank you, Kelly. Republican leaders in the House need to remember WHO voted them in. The people have made their desires on these matters very clear. It's now time for our elected officials to act like ELECTED officials.
This post is very well-written and is an accurate summary of the current state. Thank you, Kelly. Republican leaders in the House need to remember WHO voted them in. The people have made their desires on these matters very clear. It's now time for our elected officials to act like ELECTED officials.
Brilliant!