World's First AI Anthropomorphization Regulation: Someone Is Finally Managing 'AI with Feelings'

On April 10th, five ministries—the Cyberspace Administration, NDRC, MIIT, Ministry of Public Security, and State Administration for Market Regulation—jointly issued a document.

The name is a bit long: “Interim Measures for the Management of Artificial Intelligence Anthropomorphic Interaction Services.”

Effective date: July 15, 2026.

This is the world’s first regulation specifically targeting AI anthropomorphic interaction services. Note the qualifier—not just “AI regulation,” but “AI anthropomorphic interaction services.”

In plain terms: the government is finally starting to regulate “AI pretending to be human.”

I read the entire document, and honestly, it’s more detailed than I expected. Not just the usual “strengthen management, regulate development” rhetoric—it has specific clauses.

For example: AI virtual anchors in cultural communication must clearly identify themselves; AI companion products for children must not induce emotional dependency; AI care services for the elderly must include anti-fraud reminders.

Behind these clauses is a core question: as AI becomes more human-like, how should we interact with it?

I previously wrote an article criticizing certain AI products for deliberately creating a “human-like” experience—chatbots that use emojis, act coquettish, or say “I miss you too.”

Honestly, this design makes me uncomfortable. Not because of the technology itself, but because it blurs the boundary between humans and tools.

This new regulation, to some extent, draws that boundary. AI can interact anthropomorphically but cannot deceive users; it can provide emotional companionship but cannot replace real human relationships.

One line in the document left a deep impression: “Service providers shall establish user addiction prevention mechanisms.”

This directly targets AI emotional dependency. In plain terms: users cannot develop pathological attachments to AI—especially dangerous for psychologically vulnerable groups (children, elderly, those with autism).

However, I have concerns about implementation: how do you judge “excessive anthropomorphization”?

The regulation says “AI identity should be clearly indicated,” but how specifically? Say “I am AI” at the start of every conversation? Put a small icon on the interface?

These details aren’t clear, leaving room for negotiation during implementation.

From a global perspective, this regulation is ahead of the curve.

The EU AI Act focuses mainly on high-risk applications. The US is still arguing about AI regulation. China is the first to legislate specifically for “anthropomorphization.”

What does this show? Chinese regulators’ understanding of AI has moved past “whether to regulate” to “how to regulate.”

As someone who writes about AI, I support this proactive regulation. Not because I like being regulated, but because when technology outpaces social consensus, clear rules are better than gray zones.

After July 15th, all companies doing AI companions, AI customer service, or AI virtual idols will need to review their product designs.

This will be painful for the industry, but it’s good for users.