Human-Centered AI Guidance for K-12 Public Schools (Version 3.0)

Washington OSPI's statewide K-12 AI guidance requires a 'Human inquiry — AI — Human empowerment' (H AI H) frame — students and educators must open with human inquiry and close with human reflection, edits, and understanding on every AI use.

Washington Office of Superintendent of Public Instruction (OSPI) — Building AI Foundations: A Human-Centered Approach, Version 3.0 (July 1, 2024). Introduces a human-centered approach to using AI in K-12 education organized around the 'Human inquiry — AI — Human empowerment' framework, abbreviated 'H AI H': uses of AI in schools should always start with human inquiry and always end with human reflection, insight, and empowerment.

Message from Superintendent Chris Reykdal. AI is emerging rapidly across teaching, learning, and district operations. Washington is positioned to integrate AI with excitement and appropriate caution; slowing down is not an option because students and educators are already using AI. Empowerment of critical thinking is the guiding question. Teachers must be equipped with resources, training, and support so AI enhances instruction and nurtures student critical thinking. Start with human inquiry, see what AI produces, and always close with human reflection, human edits, and human understanding.

Executive Summary — In This Guidance. Understanding AI (context and framework for public education), Definitions (what generative AI is and is not, opportunities and risks), Principles and Values (equitable and inclusive, safe and secure, understandable, purposeful and beneficial), Guidance (human-centered learning environments, implementing AI in student learning, protecting sensitive and confidential data), Policy (acceptable use, requiring human input and review of AI outputs), Academic Integrity and AI Assistance (honest use, citing AI, when AI use becomes plagiarism), and Professional Development.

How to Use This Guidance. School district administrators integrate the guidance into district and school AI-use policies; educators use it for classroom implementation; students and families use it to understand suggested uses and Washington's statewide approach. The guidance is designed to evolve with advances in AI and adapt to the unique needs of Washington school communities.

Understanding Artificial Intelligence. AI is not a replacement for human intelligence or humanitarian presence in education. LEAs should build on existing policies grounded in educational integrity, student safety, and proven instructional practice — AI policy should not be written separately because AI touches student data privacy, plagiarism, cyberbullying, and digital literacy that are already governed by existing policies. State and local policymakers must build an ethical framework of funding to support AI-related policies that embrace each student's unique abilities in a safe learning environment.

AI in Education — A Human-Centered Approach. A human-centered AI learning environment prioritizes the needs, abilities, and experiences of students, educators, and administrators by: developing student AI literacy; ensuring ethical, equitable, and safe use by protecting data privacy and security, addressing biases and harms, and promoting digital citizenship; providing professional development so educators can integrate AI into pedagogy, curriculum, and assessment; applying human-centered design principles (stakeholder involvement, testing and iteration, impact evaluation); and aligning AI solutions with learner agency, collaboration, feedback, and critical thinking.

Definitions. Generative AI (Gen AI) refers to software tools trained on large amounts of data to produce text, images, videos, or other digital artifacts, including LLM-based text generation, image generation, AI tutoring, virtual assistants, and lesson-planning or grading tools. Generative AI IS: a means to augment teaching and learning; already embedded into many technologies; permeated with flaws like algorithmic bias; a product of human-led companies with their own values, agendas, and limitations; an algorithm that generates new content from a pre-trained LLM. Generative AI IS NOT: a replacement for student development; something that can plausibly be avoided or turned off; a source of unquestionable factual information; produced in a vacuum free of societal influence; a sentient being with superhuman capabilities; a replacement for highly qualified educators.

Potential Opportunities. Personalize learning and feedback in real time; lesson-plan and assessment design with customized differentiation; translation between languages; development of critical thinking through human input, data output, and elevated human analysis; aid in creativity, simulation, and skill development.

Potential Risks to Mitigate. Bias, hallucinated or inaccurate output, privacy and data-security exposure, over-reliance that undermines student development, inequitable access, and academic integrity risks.

Principles and Values. Uses of AI in Washington public education must be equitable and inclusive, safe and secure, understandable, and purposeful and beneficial.

Guidance for Implementation. Considerations for creating a human-centered AI learning environment, implementing AI in student learning through the H AI H frame, and explicitly protecting sensitive and confidential data when adopting or using AI tools.

Policy. Districts should build human-centered AI policies and update existing policies (acceptable use, academic integrity, privacy, cyberbullying, digital literacy) rather than treat AI as a standalone domain. Policies should specify acceptable use and require both human input and human review of AI outputs.

Academic Integrity and AI Assistance. Defines academically honest uses of AI, how students and staff should cite AI-generated content, and where AI use crosses into plagiarism.

Professional Development. LEAs should provide training on AI literacy, ethical and equitable use, integration into instruction, privacy and security, and evaluation of AI outputs.