China vs. the US: How Each Protects Children From AI Companions, and Why Beijing's Approach Is Stricter

China has banned "virtual intimacy" AI services — virtual partners and virtual relatives — for all users under 18. The Cyberspace Administration of China (CAC) "Interim Measures for the Regulation of Anthropomorphic AI Services" take effect on 15 July 2026. The US, meanwhile, is taking a different path, relying on patchwork state laws and stalled federal initiatives.

What China Has Banned

China has introduced a direct ban on AI companions for minors and mandatory parental consent for younger children. According to the Cyberspace Administration of China (CAC), the new rules cover any "human-like" AI services, from chatbots to virtual characters.

• A ban on "virtual intimacy" (romantic partners, virtual relatives) for everyone under 18, effective 15 July 2026
• For children under 14, explicit parental consent is required for any human-like AI interaction
• Parents receive risk alerts, can block specific characters, and can limit spending inside the service
• All AI content must be clearly labeled as machine-generated
• A separate classification system for online content aimed at minors has been in force since 1 March 2026

As broadcaster CGTN reported, the ban on virtual AI partners for minors became one of the document's most prominent provisions.

"UNICEF welcomes

China's pioneering rules protecting children from AI-related risks," the United Nations Children's Fund said in a statement.

Why the US Approach Is Stalling

The US relies on patchwork state laws and federal initiatives, which are slowed by court challenges and disputes over AI regulation. Unlike China's centralized ban, the American system is built from the bottom up and runs into constitutional limits.

At the federal level, the House of Representatives passed the bipartisan Kids Internet and Digital Safety Act (KIDS Act) package on 29 June 2026 by a vote of 267 to 117. The package includes eight bills, among them a version of KOSA and safeguards for minors interacting with AI chatbots. At the state level, Idaho, Oregon, and Washington have passed laws that prohibit chatbots from claiming to be conscious and from engaging in sexual conversations with minors.

Many state restrictions on children's access to social media and algorithmic feeds have stalled because of First Amendment lawsuits. Congress's years-long effort to make Big Tech protect children has now become entangled with the debate over AI regulation, adding new obstacles to a long-overdue deal.

What This Means

The world's two largest economies have chosen opposite models for protecting children from AI: China a fast, centralized ban, the US slow negotiation through the courts and the states. According to Bloomberg's assessment, Beijing's proactive approach to regulating AI and social media looks more effective so far, although the price of that speed is tight state control. For parents and AI service developers, this means the rules of the game diverge across the two key markets: in China, restrictions are spelled out in advance; in the US, they form a constantly shifting mosaic of state laws.

Frequently Asked Questions

What exactly do China's new rules ban?

The rules ban "virtual intimacy" AI services (romantic partners and virtual relatives) for everyone under 18, and for children under 14 they require explicit parental consent for any human-like AI interaction.

When do the rules take effect?

The "Interim Measures" on anthropomorphic AI services take effect on 15 July 2026. A separate classification system for online content aimed at minors took effect earlier, on 1 March 2026.

What is the US doing to protect children?

The US is acting through state laws and the federal KIDS Act, which the House of Representatives passed on 29 June 2026 (267 to 117). Idaho, Oregon, and Washington have restricted the behavior of AI chatbots, but many measures are held up by First Amendment lawsuits.