Congress Must Not Undermine State Efforts To Regulate AI Harms to Children

A cornerstone of conservative philosophy is that policy decisions should generally be left to the states. Apparently, this does not apply when the topic is artificial intelligence (AI).

In the name of promoting innovation, and at the urging of the tech industry, Congress quietly included in a 1,000-page bill a single sentence that has the power to undermine efforts to protect against the dangers of unfettered AI development. The sentence imposes a ten-year ban on state regulation of AI, including prohibiting the enforcement of laws already on the books. This brazen approach crossed the line even for conservative U.S. Representative Marjorie Taylor Greene, who remarked, “We have no idea what AI will be capable of in the next 10 years, and giving it free rein and tying states' hands is potentially dangerous.” She’s right. And it is especially dangerous for children.

We are already beginning to see the consequences for our children of the uninhibited, rapid, and expansive growth of AI. One clear example is the proliferation of deepfake nudes— AI-generated images that depict real people in sexually explicit scenarios. Too often, these “real people” are children. A recent survey revealed that 1 in 8 teens report knowing a peer who has been the target of deepfake nudes. The American Academy of Pediatrics warns that these child victims can experience emotional distress, bullying, and harassment, leading to self-harm and suicidal ideation.

AI is also being used to create pornographic images of real children to share in pedophilic forums or exploit children in “sextortion” schemes. In 2024, the national CyberTipline received more than 20.5 million reports of online child exploitation, representing 29.2 million separate incidents. Each of these incidents involves images that can be shared over and over. The initial harm can be devastating, and the continued trauma unbearable.

Chatbots present another alarming threat. From a 9-year-old child exposed to “hypersexualized content” to a 17-year-old encouraged to consider killing his parents, these AI-powered companions are emotionally entangling children at the expense of their mental health and safety. The American Psychological Association (APA) has expressed “grave concerns” about these unregulated technologies. The APA cites the case of a fourteen-year-old Florida boy who had developed an “emotionally and sexually abusive relationship” with an AI chatbot. In February 2024, he shot himself following a conversation in which the bot pleaded with him to “come home to me as soon as possible.” The current lack of safeguards around AI has life-and-death consequences.

Despite widespread concern about the risks of AI, there is still no comprehensive federal framework governing it. While the technology evolves at breakneck speed, federal policymakers are moving at a glacial pace. That is why much of the work to protect children has been done by state legislatures. Many states—both red and blue—have stepped up. California and Utah have passed laws to limit algorithmic abuse, require transparency, and provide innovative legal tools to protect children online. This year, states as diverse as Montana, Massachusetts, Maine, and Arizona have introduced, and in some cases already enacted, provisions to protect children from AI-related harms. These are not fringe efforts. They are practical, bipartisan attempts to regulate an industry that has demonstrated, time and again, that it will not effectively police itself.

Despite these bipartisan state efforts, Congress appears poised to halt and undo all progress aimed at keeping children safe. On June 5, Senate Republicans, recognizing that the original ban likely wouldn’t survive Senate rules, got creative. Instead of an outright moratorium, their version ties access to critical broadband funding to a state's willingness to halt any regulation of AI. That means states trying to shield children from AI-driven harm could lose out on the infrastructure dollars needed to connect underserved communities, like low-income and rural communities, to high-speed internet. It’s a cynical use of power: forcing states to choose between protecting children and connecting their most vulnerable communities to a vital resource.

Congress must abandon its pursuit of pleasing tech companies at the cost of child safety. At a minimum, Congress should strike this harmful, deeply flawed provision from the reconciliation bill. Children’s lives depend on it. If Congress wishes to play a constructive role, it should work toward setting a federal floor of protection while preserving states’ authority to go further. Very often, the best solutions to national problems come from experimentation and innovation within states. This is especially likely to be true in the complex and often confounding realm of emerging and rapidly developing technology. Allowing states—the “laboratories of democracy”—to take bold action to address the concerns of parents, children, and their communities may be the most efficient and effective way to make progress. We need Congress to work alongside and learn from state lawmakers in this endeavor, rather than standing in their way.

Jessica K. Heldman is a Fellmeth-Peterson associate professor in child rights and Melanie Delgado is a senior staff attorney at the Children’s Advocacy Institute at the University of San Diego School of Law.