“AI + education” is moving from classroom pilots to regional system design. Chongqing's Nan'an District education authority has released ten measures for deepening “AI + education” in primary and secondary schools, setting a 2030 goal of “five ones”: one regional education dataset, one cadre of hybrid-skilled teachers, a group of innovation-focused schools, one signature AI scenario per school — and one smart study companion for every student — on the way to becoming a demonstration zone.
1. Hard class-hour floors: AI literacy for all
The measures set explicit AI-literacy hours: at least 10 class hours for grades 1-2, 25 for grades 3-6, and 30 for secondary school, within a three-tier system of awareness, applied practice and innovation, supported by universities and research institutes. On the platform side, the district will aggregate education data via Chongqing's integrated public data platform, build a regional multimodal corpus, explore an education-specific LLM, equip all schools with smart teaching devices, pilot “education computing vouchers”, and aim for “100% of students experiencing 100% of experiments”.
2. One AI teacher per six classrooms
The most closely watched provision: by 2030, one dedicated or part-time AI teacher for every six classrooms, plus 5 city-level and 100 district-level seed teachers radiating training to more than 5,000 educators. Students' “smart companions” get personal learning spaces with per-user access, powered by recommendation, learning analytics and mistake-coaching; teachers gain AI tools for learning assessment, co-planning, grading and teaching-data analysis.
3. Guardrails against “AI doing the homework”
Addressing the biggest parental worry, the district's education LLM will in Q&A scenarios offer hints, prompts and follow-up questions — never direct answers. Three hard lines: no using AI to complete entire assignments, no generating harmful content, no unsupervised use. The plan also explores an “AI + brain-science learning lab”, “digital twins of scientists”, and at least one AI study trip per school each semester.
District officials admit the per-student companion “tests our platform, computing power and teaching force”, but kept the goal anyway: “It is the most direct way AI can help every student.” The target is a demonstration zone leading western China.
With class-hour floors, staffing ratios and usage red lines, this district-level plan offers a decomposable implementation framework — echoing Guangdong's WorkBuddy education edition (100,000+ teacher sign-ups) and iFlytek's Spark X2.6 teacher assistant. NineZenith's Tianxing LLM brings the same foundation to education: private deployment and knowledge-base capabilities for regional education platforms.
(Compiled from Jiemodui and public materials of Nan'an District Education Bureau.)