🧠 Food for thought

Implications, context, and why it matters.

The new AI rules go beyond content filters

• The regulations start on July 15, giving companies time to adjust products.
• The requirements go past bans on harmful material. They add operating duties, including reminders after two hours of continuous interaction and required security assessments for services with over 1 million registered users or more than 100,000 monthly active users 1.
• Providers cannot use interaction data, including sensitive personal information, to train models without separate consent. Extra safeguards apply to minors, which can slow model improvement 1.
• The policy aims at a fast-growing domestic market for AI social interaction apps 2.

China’s AI companion rules create a global test case for digital well-being

• The framework moves past content moderation and reaches into how AI companions are built and run.
• It requires usage “circuit breakers,” meaning prompts or limits that interrupt extended use, plus easy exit options. These constraints cut against engagement-maximizing design patterns common in tech 1.
• A 2026 study of Chinese users linked higher usage frequency with stronger emotional attachment. The researchers also warned about overreliance and pullback from offline relationships for some users 3.
• Rolling out these design limits at scale creates a global “natural experiment,” meaning a real-world policy test whose outcomes other governments can watch. The results could shape future AI policy in the US and EU 2.