K-12Issue 555 records
K–12
Howard County, Maryland, adopted an AI policy a week after the state deadline that makes every assignment AI-free unless a teacher allows it, joining San Mateo Union and Highland, Illinois, in putting the decision with teachers per assignment. Montclair, New Jersey, says its new pre-K-8 student AI ban and device limits will take time to implement. An EdWeek Research Center survey finds nearly a third of district and school leaders use no generative AI for operations, with communications and translation leading among those that do.
- Evidence records
- 5
- Cross-source patterns
- 3
- Evidence classes
- 4 independent reporting1 independent research
- Outcomes
- 3 emerging1 cautionary1 mixed
- Source freshness
- 5 new this fortnight
- Research completed
- 2026-11-02
Choose a role to see its takeaway beside every record in the ledger.
Patterns across the evidence
Synthesis · Lighthouse Advisory interpretation
District policies are converging on AI off by default, switched on by the teacher per assignment
Howard County makes every assignment AI-free unless a teacher allows it, SMUHSD requires teacher permission plus a record of use, Highland has teachers pick one of five levels per assignment, and Montclair allows high school use only when a teacher says so. The design moves the key decision to individual teachers, which raises consistency and training needs.
Operating questionHow will a student know, for each assignment, what AI use is permitted, and how is that applied consistently across teachers?
Supporting evidenceHoward County Public School System (reported by the Baltimore Sun via GovTech/TNS)San Mateo Union High School District (reported by The San Mateo Record)Highland School District, Illinois (reported by The Troy Times Tribune)Montclair Public Schools (reported by Montclair Local)
Adoption is outrunning implementation capacity
Howard County adopted its policy after Maryland's deadline, Montclair says compliance will take time and staff may not be fully there, SMUHSD cannot confirm which planned artifacts are done, and EdWeek's survey finds leaders cite missing staff skills as a barrier to operational AI. Policies exist faster than the lists, training and processes that make them work.
Operating questionWhich implementation artifacts (approved-tool list, training, exceptions process) exist today, and who owns finishing each?
Supporting evidenceHoward County Public School System (reported by the Baltimore Sun via GovTech/TNS)Montclair Public Schools (reported by Montclair Local)San Mateo Union High School District (reported by The San Mateo Record)Education Week / EdWeek Research Center
Detecting AI use remains unresolved
Howard County board members wanted training so teachers do not incorrectly accuse students of AI use, while Highland teachers use AI checkers to help identify authentic work. Neither source reports detector accuracy or safeguards.
Operating questionWhat role, if any, may detector output play in academic-integrity decisions, and what evidence is required beyond it?
Supporting evidenceHoward County Public School System (reported by the Baltimore Sun via GovTech/TNS)Highland School District, Illinois (reported by The Troy Times Tribune)
Evidence ledger
Full record · every source keeps its link and limitations
Howard County (MD) board adopts AI policy a week after the state deadline: assignments are AI-free unless a teacher allows it, and staff may not use AI for grading or counseling
Howard County's Board of Education approved an AI policy on Thursday (Oct 29), a week after the Oct 22 deadline set by Maryland's Artificial Intelligence Ready Schools Act; Howard is one of four Maryland systems that missed it, with Anne Arundel, Baltimore County and Baltimore City. Under the policy, effective immediately, all assignments are AI-free unless a teacher allows use, with disclosure or citation required. It sets nine standards including data privacy, human oversight, and procurement and vendor governance. Staff may use AI for lesson planning, drafts and translation but not grading, counseling or entering personally identifiable information. Howard gave all high schoolers Gemini access in January.
Why it matters, evidence and limitations
- Why it matters
- Newly current (Oct 29 vote, Oct 30 report) and directly K-12: it shows a state mandate driving district policy on a deadline, and a district that already deployed Gemini choosing teacher-gated use rather than the student ban adopted by neighboring Frederick County.
- Evidence and measured results
- News report of a board vote with quotes from district staff and board members. No outcome data; the policy text itself was not opened in this run.
- Limitations and uncertainty
- Secondhand news account of the vote; the policy text and implementation procedures were not inspected. Board members raised unresolved questions about family opt-out and training to avoid wrongly accusing students of AI use.
EdWeek Research Center: nearly a third of district and school leaders use no generative AI for operations; communications and translation lead among those that do
In a nationally representative survey of 852 educators, including 262 district and school leaders, conducted Aug 26 to Oct 6, 2026, nearly one-third of leaders said they do not use generative AI for any operational purpose. Among users, the top five uses were communications, foreign-language translation, strategic planning, employee evaluation and class scheduling. The National Principals Association CEO tied the communications focus to student privacy caution; leaders cited a lack of staff who can use the technology effectively and an absence of national guidelines. Peninsula School District (WA) uses AI, including agents, for routing, enrollment reporting and substitute-rate analysis.
Why it matters, evidence and limitations
- Why it matters
- Current (Oct 30) national data on the non-instructional side of K-12 AI, where adoption is thinner and focused on communications, which helps size realistic operational AI work in districts.
- Evidence and measured results
- Survey results reported by EdWeek's own research center with sample size and field dates; the full questionnaire, margins of error and percentages beyond the top-five ranking were not shown in the article.
- Limitations and uncertainty
- Article reports rankings, not percentages, for each use; leader subsample is 262. Peninsula's uses are anecdotal and unmeasured.
San Mateo Union High School District (CA) updates trustees on AI classroom guidance: teacher permission, use records and disclosure, with a 2026-27 AI Steering Committee
SMUHSD (about 9,000 students, six comprehensive high schools) updated trustees Oct 22 on AI guidance developed since a May 14 study session and established a 2026-27 AI Steering Committee. Published guidance treats AI as a learning aid: students must get teacher permission before using generative AI for assigned work, keep a record of use and acknowledge the tool. Inappropriate uses include undisclosed AI submissions, evading plagiarism detection and using unverified AI output; students and staff are advised not to enter personally identifiable information. May recommendations included a red-yellow-green framework, an approved platform list and short student lessons targeted for August 2026, but the district summary does not establish which were completed.
Why it matters, evidence and limitations
- Why it matters
- Current (Oct 30) example of a mid-sized high school district formalizing teacher-permission AI rules and naming year-two priorities (AI literacy, family education, assessment redesign).
- Evidence and measured results
- Local news summary of a board presentation and the district's AI guidance page; no outcome data.
- Limitations and uncertainty
- The article itself says the district summary does not establish which May recommendations were completed. District guidance page not opened in this run.
Montclair (NJ) technology town hall: after adopting a pre-K-8 student AI ban and Chromebook limits, district says implementation will take time and staff may not yet be fully compliant
At an Oct 28 technology town hall, Montclair leaders discussed implementing policies adopted Oct 14: a generative AI policy barring pre-K-8 student use except narrow, superintendent-approved, research-supported exceptions, and high school use only when a teacher permits, plus Chromebook and edtech limits for K-8. The district said a technology plan will include an annually reviewed list of approved programs checked for safety, privacy and security; Chromebook limits may take time and staff may not be fully compliant, partly because fall benchmark assessments run on Chromebooks. Parents split between calls for a total AI ban and arguments for teaching students to use and check AI.
Why it matters, evidence and limitations
- Why it matters
- Newly current (Oct 28-29): shows the gap between adopting restrictive AI policy and operating it when assessments and curriculum still depend on devices, a constraint other restrictive districts will face.
- Evidence and measured results
- Local reporting of a public town hall with named officials and parents; no outcome data.
- Limitations and uncertainty
- Policy details for high school exceptions are still being decided; the article reports the superintendent citing 'overwhelming evidence' of harm without specifying studies.
Highland (IL) school district uses a five-level assignment AI framework after Illinois state guidance, gives teachers MagicSchool, and relies on AI checkers to identify authentic work
Highland's board approved an AI policy framework Sept 28 following Illinois State Board of Education guidance issued July 9 under Public Act 104-0399. At Highland High School teachers set one of five levels per assignment, from no AI to full integration. AI use is concentrated in grades 6-12, who have take-home Chromebooks; the assistant superintendent acknowledged verifying independent work at home is harder. Teachers have MagicSchool and training, and use AI checkers to help identify authentic work. Staff are told not to enter student names, birth dates or identifying details into AI tools, and the anti-bullying policy now covers AI replicas and fabricated content.
Why it matters, evidence and limitations
- Why it matters
- Newly reported (Oct 29) example of state guidance translated into assignment-level rules; its reliance on AI checkers contrasts with other district policies warning that detectors are unreliable.
- Evidence and measured results
- Local reporting with district administrator quotes; no outcome or detector-accuracy data.
- Limitations and uncertainty
- Single-source local news account; no evidence on how AI checker results are used in discipline or their accuracy.
How to read this edition
Source findings, measured results and limitations come from the cited publications. Patterns, operating questions, role takeaways and implementation considerations are Lighthouse Advisory interpretation, stated as questions to validate locally rather than guaranteed outcomes. Vendor and operator claims are labeled as claims. Full research method.
- Independent reporting
- Independent reporting with attributable sources but without a formal evaluation design.
- Independent research
- Research conducted outside the implementing organization.