A district curriculum director is putting together next year's strategic plan. She knows she needs an "AI literacy" line item but is not sure what it actually has to contain. She wants something more rigorous than "teach students to use ChatGPT," and she wants whatever the plan says to still be defensible in 2029, when the rest of the world will be measuring this directly.

AI literacy now has a concrete global benchmark. PISA — the Programme for International Student Assessment, the OECD's triennial cross-country measurement of fifteen-year-old students — will include Media and Artificial Intelligence Literacy as an innovative domain in its 2029 cycle. The OECD itself (the Organisation for Economic Co-operation and Development, an intergovernmental body of thirty-eight mostly high-income member countries that publishes shared policy benchmarks) describes the assessment as a way to understand whether students have had opportunities to engage proactively and critically in a world shaped by media platforms and AI systems.

For school leaders, that matters because AI literacy is no longer an optional enrichment topic. Students will need to evaluate digital content, understand how AI systems mediate information, act ethically with generated media, and participate responsibly in AI-shaped spaces.

UNESCO's teacher AI competency framework points in the same direction for adults. It organizes teacher learning around a human-centered mindset, ethics of AI, AI foundations and applications, AI pedagogy, and AI for professional learning. Those dimensions can become a staff-development spine.

Five shifts deserve planning attention. Written as questions rather than promises, they give school leaders something to act on instead of something to predict.

Predictive analytics can help identify patterns earlier, but it can also label students prematurely. Edtech personalization can adapt practice and feedback, but it can narrow the curriculum if optimization replaces teacher judgment. Virtual learning environments can support community and immediate help, but they can also monitor students too aggressively.

AI tutoring and mentoring can expand access to help, but students still need human relationships and real accountability. Administrative efficiency can reduce workload, but schools must decide which tasks should remain human because they communicate care, trust, or professional responsibility.

Labor-market shifts matter. The World Economic Forum's Future of Jobs research has repeatedly named AI literacy and analytical thinking among the fastest-rising skill demands, and that signal belongs in any future-readiness conversation. But education is larger than job-market adaptation, and schools that reduce themselves to a workforce pipeline tend to produce graduates who can pass tests and not much else.

A better framing is durable agency. Students need to know how to use AI tools, but also how to read deeply, write clearly, reason from evidence, explain their process, collaborate with people, recognize manipulation, and make ethical judgments when tools make bad choices easy.

Matt Miller asks the right framing question in AI for Educators: "Plan instruction with this question front and center: Does this prepare my students for their futures?" Before adopting the next tool, leadership teams should be able to answer the prior question — what do we want from our students, and what do we want for them?

Every AI-supported task can end with a short reflection: What did the tool help you do? What did you have to decide yourself? What did you check? What did you reject? What would you be unable to explain without the tool?

That habit protects learning because it turns AI use into metacognition. Students practice seeing the difference between output and understanding, assistance and authorship, speed and quality, convenience and judgment.

Policy & Ethics

Boundaries, detectors, and school decisions

AI Ethics

Frameworks for ethical action

Educator Scenarios

Pressure-test the policy