{"resourceId":"seattle-times-wa-ai-companion-safeguards-20260918","versions":[{"version":"external-eb1ea8c9a9f751f8830643e6a926de98446b6f9cbd3752bd9b89319d7b61139e","resource":{"id":"seattle-times-wa-ai-companion-safeguards-20260918","title":"Seattle Times: Washington companion-chatbot youth safeguards contrast SPS enablement with NYC/LA pauses","organization":"The Seattle Times","sector":"Independent journalism (mental health / education)","geography":"Washington state (with U.S. megadistrict comparison)","publishedAt":"18 September 2026","publicationDate":"2026-09-18","eventDate":null,"sourceName":"The Seattle Times","sourceLabel":"Independent reporting on WA AI youth safeguards and school approaches (HTML)","sourceUrl":"https://www.seattletimes.com/seattle-news/mental-health/keeping-kids-safe-in-the-age-of-ai-what-safeguards-does-wa-have/","evidenceClass":"independent-reporting","outcomeClass":"emerging","topics":["data-security","governance-procurement","accessibility-workforce","operating-model"],"finding":"Sep 18 Seattle Times reports Washington’s spring 2026 law requiring companion chatbots to notify minors they are interacting with AI at the start and at least once per hour; use reasonable measures against manipulative engagement and sexual content with minors; disclose non-healthcare-provider status for health advice; and create self-harm/suicidality detection with crisis referrals. Rules take effect Jan 1 and are enforceable via the Consumer Protection Act. The piece contrasts Seattle Public Schools’ ethical-AI enablement (tools such as MagicSchool AI, Canva, Adobe Firefly) and OSPI human-centered guidance with NYC’s yearlong GenAI limits for elementary/middle grades plus all-grade companion bans and LAUSD’s temporary GenAI block on student devices. Also notes OpenAI ChatGPT for Teens protections and Microsoft-AFT school AI safety/privacy standard. Why now: complements 2026-09-23 archived megadistrict-pause coverage with a state-level companion-safety control path.","sledRelevance":"Gives SLED districts a concrete companion-chatbot control checklist and shows enablement and pause can coexist with youth mental-health safeguards. Relevant to developmental-risk evaluation in the K-12 brief.","evidence":"Journalistic synthesis of enacted WA requirements, district website statements, and national comparisons; not a statute annotated code extract or outcome evaluation.","architectureImplications":"If companion or social chatbots are present, require age-aware disclosure cadence, content filters, and crisis-referral hooks; separate instructional tools from companion products in the architecture.","governanceImplications":"Align school AI policies with state chatbot youth rules; document whether district tools are instructional assistants vs companion products; communicate SPS-style enablement criteria vs pause rationales to boards/parents.","securityPrivacyImplications":"Health-advice disclosures and self-harm protocols are safety controls, not substitutes for FERPA/COPPA DPAs on instructional AI; treat companion features as higher developmental risk.","caveats":"Secondary reporting; Ballmer Group funds Mental Health Project with stated editorial control retained; SPS tool inventory may change; effectiveness of chatbot crisis detection unmeasured here.","streamIds":["k12"],"roles":{"sales":"Sales framing: separate instructional AI from companion/social chatbots. This reporting shows WA imposing hourly AI-identity disclosures and crisis protocols for companions while SPS enables vetted instructional tools. Help districts map their stack to those categories and draft board language—without claiming any companion vendor meets the Jan 1 bar or that pauses are universally superior.","engineering":"Engineering framing: classify products as companion vs instructional; implement minor disclosure timers, sexual-content blocks, and crisis-referral pathways where companion features exist; keep instructional tools on approved tenants. Prerequisites include an inventory of chatbot behaviors. Proof of value: successful disclosure/crisis test cases and documented SPS-like approved lists—not engagement metrics that optimize attachment.","delivery":"Delivery framing: update AI policy for companion bans or disclosures, train staff on crisis escalation, and communicate enablement criteria. Ownership: student services/mental health co-own companion risk with IT; academics own instructional tools. Acceptance criteria: policy updated before Jan 1 where WA law applies or as voluntary standard elsewhere; crisis protocol drills completed. Risks include misclassified companion features inside edtech suites and parent confusion after NYC/LA headlines."},"roleDetails":{"sales":{"customerProblem":"Districts need youth-safety controls for chatbots while deciding between pause and guided enablement after NYC/LA moves.","stakeholders":["Student services leaders","General counsel","IT/security","Principals","Parents"],"discoveryQuestions":["Which tools have companion or social modes?","How do you disclose AI identity to minors today?","What is the self-harm escalation path?","Are you enabling SPS-style vetted tools or pausing like NYC/LA?"],"valueHypothesis":"A companion-risk classification and policy update can reduce developmental exposure without forcing a total instructional AI ban.","potentialEngagement":"2 to 4 week companion/instructional AI separation review with policy redlines and board briefing.","unsupportedClaims":["That WA rules are already met by any named vendor","That crisis detection prevents harm","That SPS tools are proven effective for learning"]},"engineering":{"fit":"Instructional assistants under approved lists; companion products only with age disclosures and crisis protocols—or blocked.","architectureIntegration":"Feature flags for companion modes, hourly disclosure events for minors, content filters, and crisis-referral deep links.","prerequisites":["Product behavior inventory","Legal read of applicable state rules","Student-services protocol"],"constraints":["Journalism not statute text","Cross-state variance","Rapid vendor changes"],"security":"Log disclosures and crisis handoffs; minimize retention of sensitive chats; separate instructional telemetry from companion dialogs.","proofOfValue":["Disclosure timer tests","Crisis-referral tabletop","Approved instructional tool list published"]},"delivery":{"work":"Classify tools, update policies, train staff on companion risks, and publish parent-facing explanations of enablement vs pause.","dependencies":["Counsel","Student services","IT filtering","Communications"],"ownership":"Student services and counsel for companion safety; CAO for instructional enablement.","skillsAdoption":"Train counselors/teachers on AI-identity disclosures and when to escalate mental-health concerns off-platform.","governanceCheckpoints":["Classification complete","Policy adopted","Jan 1 readiness (WA) or voluntary equivalent","Parent FAQ published"],"acceptanceCriteria":["No unapproved companion tools on managed devices","Documented crisis protocol","Board brief comparing local path to NYC/LA"],"risks":["Hidden companion features","False confidence in automated crisis detection","Community backlash without clear communication"]}},"retrievedAt":"2026-09-25T03:11:50Z","enrichedAt":"2026-09-25T03:17:50Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Companion risks are developmental and mental-health equity issues; staff need training to escalate off-platform rather than relying on bots.","procurementImplications":"RFPs should require disclosure of companion/social features and youth-safety controls; do not bundle companion chat into instructional purchases unnoticed.","operatingModelImplications":"Maintain parallel tracks: instructional AI under academic governance; companion risk under student-services governance.","sourceVerification":{"openedUrl":"https://www.seattletimes.com/seattle-news/mental-health/keeping-kids-safe-in-the-age-of-ai-what-safeguards-does-wa-have/","referenceExcerpt":"Companion chatbots will be required to notify minors they are interacting with AI and not a human","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}