Beijing Opens Free AI Teaching-Agent Platform to Primary and Secondary Schools
The Beijing Academy of Educational Sciences has launched Zhi Xiaoyan (智小研), a general-purpose agent platform for teaching and school research, and opened it free of charge to primary and secondary schools. The platform was unveiled on 5 November at the main forum of the 10th Beijing Education Forum.
Zhi Xiaoyan was built in-house on top of large language models combined with an education-specific knowledge base. Rather than shipping a fixed set of tools, it lets teachers define an agent's role, knowledge scope and teaching scenario in natural language, so that the working methods of experienced researchers and the standards used to evaluate lessons can be captured and reused. The institute says the platform has been applied in more than 100 schools, with over 30,000 registered users, and that its guiding principle is to prompt thinking rather than supply answers, with the AI's level of intervention tapering as a student's competence grows.
Three policy documents were released at the same event. The 2026 edition of Beijing's guidance on AI in education covers five sectors — basic education, higher education, vocational education, lifelong learning and special education — plus two cross-cutting topics, AI literacy and ethics and safety, and sets out how schools should implement AI and where the boundaries lie. A separate AI literacy framework for school students is organised around three domains: understanding AI, collaborating with and creating with AI, and governing AI, arranged in three progressive tiers from foundational to integrated to excellent. A third document, a teaching guide for AI literacy courses, addresses progression from primary to senior secondary, along with assessment and safety.
Beijing also used the forum to extend its network beyond the capital. The academy is leading a new metropolitan AI-in-education innovation community whose first members include education research institutes from Tianjin, Xiong'an New Area, Shenzhen and Chengdu. The grouping is meant to work on a template of one scenario per city, one research topic per institute and one plan per school, with the aim of producing models that can be copied elsewhere. Separately, the municipal education commission and development and reform commission, together with the science and technology commission and the Zhongguancun administrative committee, started an industry-academia-research alliance attached to the national AI pilot-testing base, intended to move AI teaching tools from trial validation to scaled deployment.
The announcements come as AI literacy moves into the compulsory curriculum. Beijing has required primary and secondary schools to run AI classes since the autumn 2025 term, and the new framework and guidebook are aimed at giving those courses a measurable standard and a progression path, part of a broader effort to shift school AI courses from simply existing to being taught well.
Zhi Xiaoyan was built in-house on top of large language models combined with an education-specific knowledge base. Rather than shipping a fixed set of tools, it lets teachers define an agent's role, knowledge scope and teaching scenario in natural language, so that the working methods of experienced researchers and the standards used to evaluate lessons can be captured and reused. The institute says the platform has been applied in more than 100 schools, with over 30,000 registered users, and that its guiding principle is to prompt thinking rather than supply answers, with the AI's level of intervention tapering as a student's competence grows.
Three policy documents were released at the same event. The 2026 edition of Beijing's guidance on AI in education covers five sectors — basic education, higher education, vocational education, lifelong learning and special education — plus two cross-cutting topics, AI literacy and ethics and safety, and sets out how schools should implement AI and where the boundaries lie. A separate AI literacy framework for school students is organised around three domains: understanding AI, collaborating with and creating with AI, and governing AI, arranged in three progressive tiers from foundational to integrated to excellent. A third document, a teaching guide for AI literacy courses, addresses progression from primary to senior secondary, along with assessment and safety.
Beijing also used the forum to extend its network beyond the capital. The academy is leading a new metropolitan AI-in-education innovation community whose first members include education research institutes from Tianjin, Xiong'an New Area, Shenzhen and Chengdu. The grouping is meant to work on a template of one scenario per city, one research topic per institute and one plan per school, with the aim of producing models that can be copied elsewhere. Separately, the municipal education commission and development and reform commission, together with the science and technology commission and the Zhongguancun administrative committee, started an industry-academia-research alliance attached to the national AI pilot-testing base, intended to move AI teaching tools from trial validation to scaled deployment.
The announcements come as AI literacy moves into the compulsory curriculum. Beijing has required primary and secondary schools to run AI classes since the autumn 2025 term, and the new framework and guidebook are aimed at giving those courses a measurable standard and a progression path, part of a broader effort to shift school AI courses from simply existing to being taught well.
JMDedu is a leading B2B EdTech Media in China that informs and connects business professionals through our reports, events, and information services. Our mission is to drive the advancement of education. Our network spans the education industry in China as well as globally, including entrepreneurs, investors, startups, educators, business decision-makers, and policy decision-makers.
Copyright © 2014-2026 en.jmdedu.com All Rights Reserved