District Guidance

Policy Framework for Artificial Intelligence in K-12 Education

This model board policy provides a comprehensive framework for governing the use of artificial intelligence technologies in educational settings. The policy addresses key areas including:

• Purpose and scope of AI use in education
• Guiding principles for ethical AI implementation
• Data privacy and security requirements
• Staff training and professional development
• Student safety and digital citizenship
• Procurement and vendor evaluation standards
• Oversight and accountability measures

Download: View and Download Model Board Policy Template

This template can be adapted to meet your district's specific governance structure and educational objectives.

Administrative Regulation for AI Implementation and Management

This companion administrative regulation provides detailed procedures and guidelines for implementing the board policy. It includes specific operational guidance for:

• AI tool evaluation and approval processes
• Implementation timelines and milestones
• Staff roles and responsibilities
• Student and parent notification procedures
• Data handling and privacy protocols
• Incident reporting and response procedures
• Regular review and update schedules

Download: View and Download Administrative Regulation Template

This regulation provides the operational details needed to effectively implement your board policy.

AI Policies from Alaska School Districts

Learn from districts across Alaska that have already implemented AI policies. This collection showcases real-world examples of how various districts have adapted AI governance to meet their unique needs and circumstances.

Current Policy Examples:

• Anchorage School District - BP 5131.9 ACADEMIC HONESTY
• Bering Strait School District - BP 5131.9 ACADEMIC HONESTY | AR 5131.9 ACADEMIC HONESTY
• Chugach School District - BP 5131.9 ACADEMIC HONESTY
• Copper River School District - BP 5131.9 ACADEMIC HONESTY | AR 5131.9 ACADEMIC HONESTY
• Cordova School District - BP 5131.9 ACADEMIC HONESTY | AR 5131.9 ACADEMIC HONESTY
• Dillingham School District - BP 5131.9 ACADEMIC HONESTY | AR 5131.9 ACADEMIC HONESTY
• Fairbanks North Star Borough School District - BP 6161.3 Artificial Intelligence (AI) | AR 6161.3 Artificial Intelligence (AI)
• Additional district policies will be added as they become available

Does this AI tool align with your educational goals and district strategy?

Key Questions to Consider:

• How does this tool support specific curriculum standards or learning objectives?
• What educational problem does this solve that existing tools cannot?
• How will this tool enhance (not replace) teacher instruction?
• What is the expected return on investment in terms of student outcomes?
• How does this align with your district's equity and inclusion goals?

Documentation Required: Vendor should provide clear use cases, research on educational effectiveness, and alignment with pedagogical best practices.

Understanding the AI technology behind the tool

Critical Technical Areas:

• AI Model Information: What models are used (GPT, custom, etc.) and how are they trained?
• Data Sources: What data was used to train the AI and is it appropriate for K-12 education?
• Bias Mitigation: What steps has the vendor taken to identify and reduce algorithmic bias?
• Accuracy & Reliability: What is the tool's accuracy rate and how are errors handled?
• Updates & Maintenance: How frequently is the AI model updated and improved?

Red Flags: Vendors who cannot explain their AI technology, refuse to discuss bias mitigation, or make unrealistic accuracy claims.

Ensuring compliance with FERPA, COPPA, and state data privacy laws

Essential Privacy Protections:

• FERPA Compliance: Signed FERPA agreements and clear data handling procedures
• COPPA Compliance: Special protections for students under 13
• Data Minimization: Tool only collects data necessary for educational purposes
• Data Retention: Clear policies on how long data is stored and deletion procedures
• Data Location: Where is student data stored and processed?
• Third-Party Sharing: Is data shared with any third parties?

Required Documentation:

• Privacy policy written in plain language
• Data Processing Agreement (DPA)
• Security audit reports or certifications
• Incident response procedures

Evaluating implementation support and system compatibility

Implementation Considerations:

• Training Resources: What professional development is provided for educators?
• Technical Support: Quality and availability of customer support
• System Integration: Compatibility with existing LMS, SIS, and other tools
• Accessibility: Does the tool meet ADA and Section 508 requirements?
• Multi-language Support: Available in languages spoken by your student population
• Device Compatibility: Works on devices available to students

Ongoing Support:

• Response times for technical issues
• Regular training updates and resources
• User community or support forums

Protecting your district through proper legal agreements

Critical Contract Elements:

• Data Ownership: District retains ownership of all student data
• Usage Rights: Clear limitations on how vendor can use district data
• Liability & Indemnification: Vendor liability for data breaches or misuse
• Termination Clauses: Data return/deletion procedures upon contract end
• Service Level Agreements: Uptime guarantees and performance standards
• Intellectual Property: Ownership of student-created content

Financial Considerations:

• Total cost of ownership including training and support
• Pricing transparency and scalability
• Payment terms and cancellation policies

Testing AI tools before district-wide implementation

Recommended Pilot Process:

• Limited Scope: Start with 1-2 schools or classrooms
• Clear Metrics: Define success criteria before the pilot begins
• Teacher Feedback: Regular check-ins with participating educators
• Student Impact: Monitor both academic and engagement outcomes
• Technical Performance: Track uptime, speed, and user experience

Evaluation Criteria:

• Ease of use for teachers and students
• Measurable impact on learning outcomes
• Integration with existing workflows
• Technical reliability and performance
• Value for investment

Timeline: Allow at least 6-8 weeks for a meaningful pilot evaluation with comprehensive feedback collection.

Establishing the groundwork for successful AI integration

Key Activities:

• Assemble your AI policy team - Include diverse stakeholders from curriculum, IT, legal, and classroom educators
• Conduct current state assessment - Inventory existing technology, policies, and staff readiness
• Develop initial policy framework - Create draft policies aligned with district values and educational goals
• Stakeholder engagement - Hold listening sessions with teachers, parents, and students
• Budget planning - Identify funding sources and create implementation budget

Deliverables:

• AI implementation team roster and charter
• Current state assessment report
• Draft AI policy framework
• Stakeholder feedback summary
• Implementation budget and timeline

Testing approaches with limited scope to learn and refine

Implementation Activities:

• Launch pilot programs - Select 2-3 schools or departments for initial implementation
• Professional development rollout - Begin AI literacy training for pilot participants
• Vendor pilots - Test selected AI tools with small groups
• Policy refinement - Update policies based on pilot learnings
• Feedback mechanisms - Establish regular check-ins and data collection

Success Metrics:

• Teacher confidence and comfort with AI tools
• Student engagement and learning outcomes
• Technical performance and reliability
• Policy effectiveness and clarity
• Cost-effectiveness of selected tools

Critical: Document lessons learned and best practices for scaling in Phase 3.

District-wide rollout with continuous monitoring and improvement

Scaling Activities:

• District-wide policy rollout - Implement refined policies across all schools
• Comprehensive training program - AI literacy for all staff and students
• Technology deployment - Roll out approved AI tools district-wide
• Support system activation - Launch help desk, peer support networks, and resources
• Communication campaign - Keep all stakeholders informed of progress and changes

Monitoring & Evaluation:

• Regular policy reviews - Quarterly assessment of policy effectiveness
• Performance tracking - Monitor academic, engagement, and operational metrics
• Stakeholder feedback - Ongoing surveys and focus groups
• Technology monitoring - Track usage, performance, and security
• Continuous improvement - Regular updates based on data and feedback

Essential strategies for managing the human side of AI implementation

Building Buy-in:

• Start with willing participants - Identify and support early adopters
• Showcase quick wins - Highlight early successes and positive outcomes
• Address concerns directly - Provide clear answers about job security and student privacy
• Provide adequate support - Ensure training and help are readily available
• Celebrate progress - Recognize and reward successful implementation

Common Challenges & Solutions:

• Resistance to change: Focus on benefits and provide hands-on training
• Technology anxiety: Offer peer support and step-by-step guidance
• Privacy concerns: Provide transparent information about data protection
• Workload increases: Phase implementation to allow time for adjustment
• Inconsistent usage: Create accountability measures and support systems

Deepfakes are AI-generated or AI-manipulated media-images, audio, or video-that can convincingly depict real people saying or doing things they never actually did. Synthetic media is the broader category that includes any content created or substantially altered by AI systems. Unlike adult offenders, minors often create deepfakes out of curiosity, or as a modern form of teasing, bullying, and harassment-often failing to grasp the gravity of the harm they cause.

Types of Synthetic Media Abuse in Schools:

• Face-Swapping / Morphed Imagery: Algorithmic superimposition of a target's face onto an explicit or compromising body, often sourced from social media or school directories
• "Nudify" Applications: Apps that use neural networks to computationally strip clothing from a standard photograph, generating fabricated explicit imagery
• Audio Cloning: Brief vocal samples used to replicate a person's voice with exact inflection-increasingly used to fabricate recordings of administrators or educators making offensive statements
• Fully Synthetic Imagery: Media generated entirely by text-to-image AI models that can be engineered to resemble specific individuals

Impact on Victims:

The emotional and psychological toll on targeted students is severe. Victims experience profound trauma, anxiety, and reputational injury. They are often terrified that the broader community will believe the fake images are real and may hesitate to come forward out of fear of how widely the images have circulated. The distribution of nonconsensual images can have long-lasting, serious mental health harms on the target of abuse.

Key Terms:

• Deepfake: AI-generated media that realistically depicts a real person's likeness, voice, or actions
• Synthetic Media: Any media content (text, image, audio, video) that is generated or significantly modified by AI
• Non-Consensual Intimate Imagery (NCII): Sexually explicit images created or shared without the subject's consent, including AI-generated versions
• CSAM: Child Sexual Abuse Material - AI-generated explicit imagery of minors is legally classified as CSAM

Districts should review and update existing policies to explicitly address AI-generated synthetic media. The ICAC Task Force and Stanford HAI recommend schools take four key steps: understand the problem, develop policies and procedures, implement a victim-centered response, and invest in prevention programming.

Acceptable Use Policies:

• Explicitly prohibit the creation or distribution of deepfake content depicting students, staff, or community members without consent
• Define consequences for misuse of AI tools to create harmful synthetic media
• Address the use of school devices, networks, and accounts for generating synthetic content

Anti-Bullying & Harassment Policies:

• Expand definitions of harassment and cyberbullying to explicitly include AI-generated or AI-manipulated content
• Establish clear reporting and response procedures for deepfake incidents
• Provide support resources for students who are targeted by synthetic media

Student Code of Conduct:

• Include specific language addressing AI-generated content in academic dishonesty, harassment, and digital citizenship sections
• Outline a tiered disciplinary response that accounts for the severity and intent of deepfake creation or distribution
• Prioritize behavioral interventions and restorative justice over purely exclusionary discipline for juvenile offenders, as recommended by Stanford HAI

Staff Policies:

• Provide guidance for educators on recognizing and responding to deepfake incidents
• Establish protocols for reporting suspected AI-generated content involving minors to appropriate authorities
• Protect staff from retaliatory digital forgery and establish rapid-response crisis communication protocols

When a deepfake incident is reported, a swift, trauma-informed response is essential. The Student Privacy Compass Deepfakes Toolkit provides a comprehensive protocol for school administrators.

Critical Guidance for Administrators:

• Treat deepfakes like physical harassment: All deepfake grievances must be treated with the utmost urgency, respect, and seriousness, on par with traditional physical harassment complaints
• Victim-Centered Response: Crisis responses should be trauma-informed, ensuring victims receive immediate support. Direct victims to the NCMEC "Take It Down" program to help remove imagery from online platforms

Restorative Practices over Purely Exclusionary Discipline:

• Avoid relying solely on long-term suspensions or immediate referral to the juvenile justice system for student offenders
• Consider restorative justice frameworks-structured reflection, targeted community repair projects, and educational programs about digital harm
• Facilitated, mediated dialogue may be appropriate, but only if explicitly consented to and desired by the victim
• For first-time juvenile offenders, Stanford HAI recommends mandating educational programs or behavioral health counseling over criminalization

Strategies for Administrators:

• Education and training: Implement training programs for students and staff regarding the responsible use of generative AI tools
• Policy review: Review existing disciplinary policies to incorporate clear guidelines and legal ramifications
• Communication with stakeholders: Maintain open communication with parents, students, and faculty regarding risks
• Collaboration with law enforcement: Establish relationships with local law enforcement for proper reporting and response
• Promoting a safe environment: Foster an environment where students feel comfortable reporting incidents

Alaska statute provides several existing legal provisions relevant to deepfake incidents in schools. Districts should ensure policies reference and align with current state law.

Relevant Alaska Statutes:

• AS 11.61.120 - Harassment in the Second Degree: Addresses publishing or distributing photographs or films showing intimate areas or sexual acts; includes electronic communications that insult, taunt, or intimidate persons under 18 (Class B misdemeanor)
• AS 11.61.123 - Indecent Viewing or Production of a Picture: Addresses the knowing production of pictures involving private exposure of intimate body parts
• AS 11.41.455 - Unlawful Exploitation of a Minor: Addresses inducing or employing a child under 18 to engage in, or producing material depicting, sexual conduct, a Class A felony or unclassified felony for repeat offenders or victims under 13

Federal Legislation:

• TAKE IT DOWN Act (S.146) - Federal legislation establishing criminal and civil penalties for the knowing publication of sexually explicit images-whether authentic or digitally forged-without consent

Note: AI-specific legislation is evolving rapidly at both the state and federal levels. Districts should monitor the Alaska Legislature and the NCSL AI Legislation Tracker for updates.

The following resources provide research, guidance, and policy frameworks for addressing deepfakes and synthetic media in educational settings.

Guidance for Schools:

• ICAC Task Force - Addressing Children's Use of GenAI to Create Deepfake Imagery - Fact sheet for school-based professionals covering the issue, impact, school response, and prevention programming
• Stanford HAI - Addressing AI-Generated CSAM: Opportunities for Educational Policy - Policy brief on school preparedness gaps, mandated reporting clarity, and behavioral intervention recommendations
• Student Privacy Compass - School Administrator's Guide to Addressing Deepfakes - Comprehensive toolkit covering incident handling, CSAM evidence protocols, restorative practices, and policy templates
• Innovate Ohio - AI Policy Development Toolkit - Structured five-step methodology for creating evidence-based AI policies grounded in district values
• ACSA - The Deepfake Dilemma: Moving Forward Strategies for School Administrators - Practical strategies covering education, policy review, stakeholder communication, and law enforcement collaboration

Research & Reports:

• CDT - In Deep Trouble: Surfacing Tech-Powered Sexual Harassment in K-12 Schools - Center for Democracy & Technology research on the prevalence of NCII in schools, and gaps in school prevention and victim support
• Common Sense Media - Deepfakes Can Be a Crime: Teaching AI Literacy Can Prevent It
• UNESCO - Deepfakes and the Crisis of Knowing
• The 19th - Deepfake AI: Kids, Schools, Laws, and Policy
• NCMEC on AI-Generated CSAM - National Center for Missing & Exploited Children's response to GAI Exploitative Imagery

Victim Support & Prevention:

• NCMEC "Take It Down" Program - Free tool to help victims remove explicit online imagery
• StopNCII.org - Free tool to help victims of non-consensual intimate image abuse