Permitted but regulated, used as a supplement, not a replacement, and retaining teacher control over classroom use.
These in summary are the main points of ISD Policy 625 on the Responsible Use of Artificial Intelligence or AI in South Washington County Schools.
Adopted earlier this fall, the policy was explained in more detail at the Oct. 8 board workshop meeting.
“As you know we passed policy 625 in August I believe, and we promised we’d come back and talk about what does this look like in practice,” Assistant Superintendent Kelly Jansen said in introducing the presentation by SoWashCo School Tech Integration Coordinator Dr. Amber Sorenson and Tech Director Bob Berkowitz.
With three documents addressing three jobs, the new AI rules are supported by the board approved policy setting what and why as well as values, guardrails, and accountability, a procedures document owned by administration and spelling out how staff apply the policy, and finally staff guidance related to practice on how the policy is used.
AI performs tasks associated with human reasoning, but in a distinct manner from what humans mean by thinking, relying on mathematical formulas and pattern recognition rather than actual humanlike thought.
As such, human oversight remains essential to ensure ethical use.
“The purpose of this policy is really to state that AI is here, people are using it, and we want to make sure that we’re using it safely, responsibly, and as a supplemental tool,” Sorenson said.
Among concerns the policy addresses are student privacy, academic integrity, and ensuring human oversight.
“It authorizes it but doesn’t require anyone to use AI,” Sorenson said. The policy split into three parts begins with human input, followed by AI, with the end result reverting to human agency for an H-AI-H use model.
“The human owns the outcome,” Sorenson said. “If you’ve used AI you know it can be wrong. And it can be very confident about being wrong.”
Vendor contracts for AI tools at SoWash Co Schools undergo an approval process, including request, review, contract, and listing. Each tool has an agreement, which factors into privacy.
“Without an agreement, assume it’s public,” Berkowitz said of AI use in general. A list of approved district technology providers can be found under the district website services tab and technology subtag, or at www.sowashco.org/services/technology#tech-providers for a direct link.
Commitments under the district’s AI policy include academic integrity and original thinking, not putting personally identifiable information, transparent use with attribution, regularly reviewed privacy controls, fair access to approved tools, training on bias and misinformation from AI, preserving human decision making, age-appropriate use, and application to all AI users in the school system.
Key terms defined in the new policy include the following:
• Personally Identifiable Information (PII): names, IDs, likeness, indirect identifiers
• Hallucinations: confident but false output
• Deep fakes: synthetic media, voice cloning
• Agentic AI: plans and takes multi step actions
• Autonomous AI: acts without human approval
• District-approved tool: reviewed and authorized
The policy defines AI tools are determined by source (closed or publicly available) and who uses it (student facing, confidential, or operational).
Expectations and use differ by age in the new policy, with use highly restricted for ages K to 3, focusing on teacher directed focus on what AI is. Grades 4 to 8 have a structured introduction to AI, emphasizing rules, guided use and critical evaluation of AI-generated information. Grades 9 to 12 will entail broader access to AI, with increased expectations for transparency, accountability, and understanding how AI models work.
Finally. AI literacy will be incorporated into digital literacy instruction, with the H-AI-H model emphasized at every grade level for the district.
The four levels of acceptable AI use include AI free (students complete assignments independently without AI assistance), AI Assisted (AI may help brainstorm and organize ideas, but students produce their own work and acknowledge AI use), AI Enhanced (AI may provide feedback on writing and organization, with students deciding which suggestions to accept), and AI Empowered (Students may use AI-generated perspectives to expand their analysis and compare ideas).
With each of the above teachers determine the appropriate level for each assignment. When AI is permitted, students must acknowledge its use. Consumer AI tools without district agreements are treated as public.
AI products must undergo a district approval process involving reviews of privacy, accessibility and technical compatibility. Vendors must disclose their data practices, deletion rights and procedures for reporting security breaches.
The policy meanwhile also addresses equal access to AI technology, with the district building upon existing programs providing devices, internet connectivity and assistive technology to students.
It identifies economic circumstances, disabilities, language barriers and geographic location as potential obstacles to equitable access.
The District Technology Advisory Committee (DTAC) will provide opportunities for families and community members to comment on AI guidance, digital literacy and other technology-related decisions.
The policy also provides for staff, student, and family training in responsible AI use, as well as planned annual reviews of data safeguards, with a three tier system for reporting violations and disciplinary procedures. Divided according to severity, examples, and response, the reporting system is divided as follows:
Tier 1 - Minor (unapproved AI tools, missing attribution, individual violations, with escalation to principal or program lead)
Tier 2 - Serious (data exposure, major academic integrity violations, deepfakes of identifiable individuals, with escalation to assistant superintendent)
Tier 3 - Critical (AI-generated intimate imagery of minors, major security breaches, distribution of deep-fakes, with escalation to law enforcement, appropriate agencies and legal counsel)
Families are notified as appropriate for minor incidents, as soon as possible for serious incidents, and as soon as legally permissible for critical incidents.
District administrators will update procedures as technology changes, while staff guidance will be reviewed yearly. AI record retention is handled like other documents under policy 307, while the district's recommended AI tools list will be updated as needed.