People's Education Press and Xiaoyuan Launch First K-12 Textbook Learning AI Agent: Authoritative Content Meets Education LLMs, Moving AI Education from Tools to Systems

星战科技团队

2026-09-12 08:47:00

47 次观看

On August 21, 2026, with support from the People's Education Audio-Visual and Digital Publishing House, Xiaoyuan Learning Machine launched the first K-12 textbook learning AI agent in Beijing. Built on authoritative textbook content and Xiaoyuan's self-developed Yuanli large model, the agent uses dual AI Companion Learning and AI Guided Learning modes to support students' self-directed study. Xiaoyuan Learning Machine becomes the first smart hardware channel to distribute this AI agent....

On August 21, 2026, with support from the People's Education Audio-Visual and Digital Publishing House, Xiaoyuan Learning Machine held a launch event in Beijing for the first K-12 textbook learning AI agent. Led by the People's Education Audio-Visual and Digital Publishing House, the agent is built jointly on high-quality textbook content resources and Xiaoyuan's self-developed Yuanli large model. Through a dual-mode learning system combining AI Companion Learning and AI Guided Learning, it delivers tailored textbook learning plans for students. Xiaoyuan Learning Machine thus becomes the first smart hardware channel to release this AI agent. The industry backdrop of this collaboration is notable. In April this year, the Ministry of Education and four other departments jointly issued the "AI + Education Action Plan," explicitly calling for the development of intelligent learning companions, the integration of education large models and AI agent tools, and the creation of themed learning scenarios. Under this policy tailwind, AI agent products centered on authoritative textbooks and designed to systematically serve learning scenarios have become a critical missing piece the education industry urgently needs to fill. Long Zhengwu, General Manager of the People's Education Audio-Visual and Digital Publishing House, summarized the answer in one phrase: a two-way convergence of content and technology.

From a product design perspective, this system aims to solve not whether AI exists, but whether children can truly learn. The AI Companion Learning mode turns static textbooks into dynamic AI textbooks, presenting texts through scenario animations, with point-and-read, repeat-after-reading, recitation assessment, vocabulary learning, and AI Q&A unfolding in sequence. The AI Guided Learning mode, meanwhile, has an AI teacher plan learning paths according to curriculum standards and textbook logic, by grade, subject, and lesson type, using guided questions to drive children to think actively. The two modes respectively address whether students are willing to learn and whether they know how to learn. In home settings, textbook learning has long faced structural difficulties: not knowing where to start, not knowing what was not understood after studying, having no one to ask when questions arise, and lacking a clear path and an ever-present guide. The deployment of education AI agents must also pass the content safety test. Throughout product development, the company has maintained a strict content review system, combined with an AI-native safety architecture, establishing a full-chain mechanism covering input identification, generation control, result review, and risk feedback. Content compliance and technical reliability are both indispensable.

The large-scale deployment of education AI agents relies on continuous computing power for model training, inference, and content generation. StarWar Technology focuses on GPU computing power platforms and the OPC Agent Collaboration Platform. Through unified computing scheduling and role-based collaboration orchestration, it helps enterprises centrally manage education large model training and inference tasks, allocate resources elastically on demand, support the stable operation of complex scenarios such as AI Companion Learning and AI Guided Learning, and move AI education practical training from isolated experiences toward replicable service systems. Pilot data show that students who continuously used this AI agent increased their mastery of Chinese language knowledge points by 33.6% and English by 24%. From finishing a text to truly understanding it, and from one-size-fits-all static expression to personalized active exploration, the coupling of authoritative published content and education AI is changing the paradigm of textbook learning. It is foreseeable that as the smart hardware ecosystem expands, such AI agents will penetrate more learning scenarios. The textbook learning AI agent brings authoritative content and education AI truly into the learning process, enabling textbooks to move from accessible to genuinely understood. Whether it can scale depends on content quality, safety systems, and long-term operations, but in terms of direction, it has already written a concrete sample for AI + education.

总结

The deep coupling of authoritative textbooks and education large models is moving AI-assisted learning from tool-based assistance toward systematic learning services. The training, inference, and content generation of education AI agents place higher demands on stable computing power and orchestration-ready agent infrastructure. For enterprises, landing education AI requires not only content and data compliance but also reliable computing power and engineering support. StarWar Technology focuses on GPU computing power platforms and the OPC Agent Collaboration Platform, providing computing scheduling and engineering support for AI education practical training and AI agent applications. According to industry analysis, as AI chips, computing infrastructure, and large language model ecosystems mature, AI education will increasingly become a systems-level race.