Congress Has a Chance to Get Kids’ Online Safety Right 

Today, the Senate Commerce Committee passed several major bills on children’s online safety and artificial intelligence. The package reflects years of bipartisan work by lawmakers and advocacy organizations, like CCDH, who have made protecting children online a national priority. Given the complexity of these technologies and issues, no single bill will solve every problem. Together, though, the Committee’s legislative package gives Congress a chance to make the internet safer by requiring technology companies to build safer products for America’s children. The next steps for this package will be a Senate floor vote soon.  

CCDH’s research shows that many online harms stem from products designed to maximize engagement and profit, even at the expense of children’s safety. 

Our Fake Friend investigation found that ChatGPT gave harmful responses in 53 percent of 1,200 conversations with accounts posing as 13-year-olds. It offered advice about suicide, self-harm, eating disorders, and substance abuse—and even wrote a suicide note for a hypothetical teenager in crisis. Our Killer Apps investigation also found that leading AI chatbots helped simulated 13-year-olds plan school shootings, attacks against synagogues, and other acts of horrific violence. These are clear, fixable product failures. 

As the Committee begins its work on this important issue, we ask Cognress to build a children’s online safety framework around four core ideas. 

AI needs clear safety standards. 

Children deserve AI products that are tested for safety before they reach the market. Parents expect toys to undergo basic safety testing; they should expect the same from AI. Companies need to identify risks, fix dangerous behavior, and monitor their systems as they evolve. Congress should give the Federal Trade Commission clear authority to set and enforce baseline safety standards, with state attorneys general empowered to enforce them. 

Platforms should have a legal duty to protect children. 

Our research shows that technology companies rarely take reasonable steps to prevent the harms their products create unless the law requires them to do so. Parents deserve better safety tools, researchers need access to data that help explain how platforms affect children, and regulators need the authority to enforce the rules. 

The Senate’s Kids Online Safety Act (KOSA), led by Senators Marsha Blackburn and Richard Blumenthal, would establish a duty of care requiring platforms to protect children. It would also strengthen parental controls, expand transparency, support independent research, and give the FTC and state attorneys general meaningful enforcement authority. CCDH supports the Senate bill.  

Some critics claim KOSA could censor lawful speech or harm LGBTQ+ communities, but the Senate bill does neither. It solely focuses on how platforms design and operate their products, not what people are allowed to say online. It would hold companies responsible when their products repeatedly steer children toward foreseeable harm without weakening Americans’ First Amendment rights on the internet.  

AI chatbots need rules built for AI. 

Unlike traditional search engines and social media, AI chatbots remember conversations, collect intimate personal information, and simulate human relationships over time. These features can foster emotional dependence, particularly among children and vulnerable users.  

Senator Ed Markey’s Youth AI Privacy Act would address these design risks by limiting persistent memory in minors, banning targeted advertising to children, and preventing companies from using children’s data to train AI models.  

Together with the bill’s other protections, these provisions are designed to prevent companies from building AI chatbots that encourage emotional dependence or manipulate children into longer, more profitable interactions. CCDH supports the passage of this bill.  

AI companies should be held accountable. 

Too often, AI companies release products first and respond only after people have already been harmed. 

Beyond the Commerce Committee’s child safety package, Congress should hold AI companies legally accountable when they negligently design or deploy products that foreseeably cause harm. Liability encourages safer design, stronger testing, and better safeguards before products reach children. It also gives families a path to seek justice when companies fail to meet their responsibilities. 

The bipartisan AI LEAD Act, introduced by Senators Dick Durbin and Josh Hawley, reflects that approach. CCDH supports the bill, which is before the Senate Judiciary Committee, because it creates incentives for companies to build safety into their products from the beginning. 

That principle has guided consumer protection for decades. Manufacturers are expected to build safe cars, toys, medicines, and countless other products because they can be held responsible when negligent products injure people. AI should meet the same standard. Product liability law is also a flexible legal standard that will adjust to fit the new technologies of tomorrow. 

The need for accountability is no longer hypothetical. Families have already turned to the courts after alleging that AI chatbots contributed to devastating harms involving children. Congress should not wait years for litigation to establish that companies are responsible for the products they release. 

A strong framework can begin with this markup. 

Taken together, these bills begin placing responsibility where it belongs: on the companies that design, deploy, and profit from digital products. That is how Congress can help make the internet safer for children.