Writing AI Into School

Part II of the AI in schools series.

The race to “teach AI” is no longer abstract policy talk—it’s being scheduled, hour by hour, onto timetables. China is moving first with compulsory class-time; the U.S. Northeast is threading AI across subjects while states stand up guardrails and training. From a teacher’s vantage, that split matters. It shapes whether AI shows up as a new prep period’s worth of content with assessments, or as a cross-curricular assistant that quietly changes planning, feedback, and the rhythm of class.

Hours on the schedule, literacy on the ledger

Beijing’s education authorities have decided that every primary and secondary student will “fully implement general AI education” starting this fall—no fewer than eight hours per academic year—and they’ll count AI course outcomes in comprehensive quality evaluations. That’s not a pilot; it’s a timetable. Courses can be taught standalone or integrated with information technology and science, and the stated focus spans concepts, applications, implementation methods, and ethics. (Beijing Municipal Education Commission regulations as reported in Global Times, 2025).

Hangzhou is going even further: at least 10 hours a year, with schools free to bunch time into a one-week intensive or weave it into tech classes. The system is explicit about a pipeline—from early curiosity to applied practice—and it pairs student learning with teacher upskilling and data protection. As SCMP summarized, local governments are acting because Beijing wants “all primary and secondary students to study” AI. (SCMP, 2025).

National guidance clarifies boundaries for classroom use. China’s Ministry of Education issued two 2025 documents: one to build a “tiered, progressive and spiraling” AI curriculum; another to set rules for generative AI—bar primary students from independently using open-ended generators and forbid teachers from substituting core teaching with AI. The aim is strong literacy without replacing the teacher’s role. (Global Times, 2025).

Learn more about AI

Stay updated with new articles about AI and prompting!

Subscribe

Email sent! Check your inbox to complete your signup.

No spam. Unsubscribe anytime.

For teachers inside that system, the message is crisp. A Beijing secondary teacher told media the new rules add AI on top of the existing curriculum and encourage practical use—“using AI to provide real-time classroom feedback and student evaluations”—while still centering teacher judgment. That quote signals two things a classroom educator hears: workload and accountability. (Global Times, 2025).

Northeast U.S.: Cross-curricular integration, with state guardrails

The Northeast story does not start with a mandate. It starts with a reversal. In May 2023, New York City’s chancellor David C. Banks acknowledged that early “knee-jerk fear and risk” about ChatGPT “overlooked the potential” for supporting teachers and students, and declared the district would “encourage and support” educators in exploring this “game-changing technology.” That reframing—from restriction to responsible use—set the tone across the region. (Banks, 2023).

Massachusetts has since published a multi-year roadmap. DESE convened a K–12 AI Task Force, created resources in 2025, and slated implementation supports for 2025–26, with policy considerations to embed AI into the state’s Curriculum Frameworks and Educator Preparation Programs in 2026–27. That sequence tells teachers: learn, try, scale—then standardize. It also signals districts and teacher-prep programs will be expected to align. (DESE, 2025).

Rhode Island released statewide guidance in August 2025 that frames AI as present-tense pedagogy, not a distant horizon. “Artificial intelligence is not the future for our schools—it’s the present,” Commissioner Angélica Infante-Green said. The document isn’t regulation; it’s a structure for ongoing conversations, paired with implementation support and an AI Advisory Group. That matters to teachers because it legitimizes classroom experiments while promising PD and feedback loops. (RIDE, 2025).

Maine’s approach is teacher-first and pragmatic. The state’s AI Guidance Toolkit pairs policy with “real-world scenarios” and explicitly tells educators to “use AI to support—not replace—human connection and critical thinking.” It also offers walkthrough sessions and positions AI within digital citizenship. For a teacher, that sentence is policy you can teach to: it gives you language for norms with students and parents. (MDOE, 2025; Maine DOE, Technology Guidance page, 2025).

New Hampshire is standardizing from the ground up, via partnerships and a collaborative framework. The NH Department of Education extended access to Khan Academy’s AI tools statewide, and state educators released a 77-page PreK–12 Generative AI Framework to guide ethical, effective use. As Ellen Hume-Howard put it, “We have to buckle down and understand it, so that we can help kids be prepared for it.” For teachers, that means vetted tools, shared expectations, and less guesswork about disclosure and assessment. (NHED, 2025; New Hampshire Bulletin, 2025; NH Learning Initiative, 2025).

Compulsory hours vs. cross-curricular diffusion

Under China’s model, teachers inherit both curriculum and compliance. There’s a set number of hours, a sequence from concepts to implementation, and evaluations tied to AI learning. That’s workload—but it’s also clarity. Teachers can plan units, anticipate assessments, and leverage new labs and partnerships with universities. It’s close to computer-science frameworks—only faster and more uniform. (Global Times, 2025; SCMP, 2025).

In the Northeast, teachers experience a softer landing. State signals say: integrate AI to improve core instruction and planning; don’t outsource judgment; build transparent policies. Surveys back up what many report anecdotally—weekly AI users are reclaiming ~6 hours a week they pour back into feedback and differentiation. “Using AI has been a game changer for me… lesson planning, communicating with parents and increasing student engagement,” said one teacher quoted by the AP. That time dividend is meaningful in classrooms where minutes are the coin of the realm. (Gallup & Walton Family Foundation, 2025; Gecker, 2025).

There’s a cautionary countercurrent: federal and civil-rights analyses warn that teachers must vet AI outputs to avoid bias and privacy pitfalls—time saved can vanish into time spent checking. Maine’s line about “support—not replace—human connection” is the practical middle. It’s showing up in district policies in Massachusetts, Rhode Island PD workshops, and New Hampshire’s framework prompts on assessment redesign and disclosure. (U.S. Dept. of Education, 2023; U.S. Commission on Civil Rights, 2024; MDOE, 2025; URI K–12 AI workshops, 2025).

Faster pilots, same guardrails

Because they control procurement and PD cycles, private and charter schools often pilot earlier. In New York City, charter leaders are pushing classroom AI strategies into practice—teacher assistants for planning, student-facing scaffolds—while navigating the same disclosure, privacy, and grading boundaries that public peers do. Their experiments flow back into district conversations via workshops, advisory groups, and vendor vetting. (Spectrum News NY1, 2025).

The policy texture teachers teach through

What ultimately shapes classroom practice isn’t a single policy—it’s the texture across guidance, training, and procurement.

• In China, teachers are handed hours, scope, and assessments. National guidelines give them cover to use tools while clearly forbidding replacement of the teacher’s role or unsupervised use by younger students. The upside is coherence; the pressure is compliance and pace. (Global Times, 2025).
• In the Northeast, teachers are handed principles and pathways. Massachusetts is building toward embedding AI in frameworks and prep programs; Rhode Island is coupling guidance with an advisory group and PD; Maine is writing teacher-friendly sentences you can put on a projector; New Hampshire is pairing access with an educator-written framework. The upside is flexibility; the risk is unevenness. (DESE, 2025; RIDE, 2025; MDOE, 2025; NHED & NHLI, 2025).

Teachers in both contexts are clear about one constant. As Banks wrote when NYC flipped from ban to embrace, the job is to ensure “benefits are equitably distributed” and that students learn to navigate a world where “understanding generative AI is crucial.” Whether the state hands you ten hours or a toolkit, the core remains deeply human: knowing your students, crafting tasks that demand thinking, and using AI to widen—not narrow—what they can do. (Banks, 2023).