AI is reshaping childhood in China (restofworld.org)

China’s rapid embrace of large AI models such as DeepSeek and Qwen is remaking childhood — from $1,170 AlphaDog robot pets (Weilan) that practice English, chat, dance and livestream home cameras, to $1,500 iFlytek tablets that generate personalized tests and grade stacks of papers in minutes. A government directive this August and provincial mandates (Beijing making AI education mandatory; Shandong equipping 200 schools and training all teachers on generative AI within 3–5 years) have accelerated deployment: AI tutors, homework-help apps, AI therapy booths (Ling Xin Intelligence), virtual mental‑health counselors, career coaches and real‑time voice chatbots (ByteDance’s Doubao) are now common in classrooms and homes. The scale has spawned a multibillion‑dollar edtech market and a rush of products promising individualized learning, companionship and 24/7 support. The significance is double-edged for AI/ML communities: technically, these deployments showcase practical uses of conversational and grading models, embedded robotics, computer vision (cameras in devices) and personalization pipelines at national scale; they create massive data and real-world feedback loops for model improvement. But researchers warn of overstated benefits — limited pedagogical change under a centralized curriculum, risks to social development, dependence on automated answers, and widening rural–urban inequities as poorer areas default to screen‑based “self‑learning.” The rapid, sometimes performative adoption highlights urgent needs for rigorous evaluation, safety standards, and research on long‑term cognitive and social impacts as AI migrates from labs into children’s daily lives.