K-12Issue 515 records
K–12
Long Beach Unified is surveying students and families about classroom AI and screens and forming a workgroup, including anti-tech advocates, to shape policy after Nov. 6. Bloomington, Minnesota approved only a small teacher pilot of Colleague AI tied to a University of Washington study of student-learning effects. A new meta-analysis shows why that measurement matters: its large average AI learning effect comes with a prediction interval that includes zero and few K-12 studies. On controls, Sandy Valley, Ohio adopted a policy on AI processing of recordings at school, and EPIC urged D.C. to write audits and data limits into school AI contracts.
- Evidence records
- 5
- Cross-source patterns
- 3
- Evidence classes
- 3 independent reporting1 independent research1 academic research
- Outcomes
- 3 emerging1 cautionary1 mixed
- Source freshness
- 5 new this fortnight
- Research completed
- 2026-10-29
Choose a role to see its takeaway beside every record in the ledger.
Patterns across the evidence
Synthesis · Lighthouse Advisory interpretation
Districts are building their own evidence and input processes before deciding AI rules
Long Beach Unified is surveying students and families and forming a workgroup, including anti-tech advocates, before recommending AI policy; Bloomington approved only a small teacher pilot paired with a university study of student-learning effects.
Operating questionBefore scaling or restricting an AI tool, what local evidence (opinion, usage, outcomes) will the district collect, and who decides what result triggers action?
Supporting evidenceLong Beach Unified School District (California), as reported by LAistBloomington Public Schools, ISD 271 (Minnesota); Colleague AI Inc.; AmplifyLearn.AI Center, University of Washington
Research averages on AI learning gains do not predict local results, which is why pilots need built-in measurement
The Frontiers meta-analysis reports a prediction interval that includes zero and likely small-study inflation, with few K-12 effects; Bloomington's pilot is structured to measure whether teacher use of Colleague AI is tied to student gains rather than relying on vendor claims.
Operating questionDoes each AI pilot have a baseline, comparison group and stop rule so the district can tell a null local effect from a positive one?
Supporting evidenceFrontiers in Education (peer-reviewed review article)Bloomington Public Schools, ISD 271 (Minnesota); Colleague AI Inc.; AmplifyLearn.AI Center, University of Washington
AI privacy controls are moving into formal board policy and procurement terms
Sandy Valley adopted a board policy specifically on recording and AI processing of communications at school; EPIC asked the D.C. Council to require pre-sale benchmarks and contract clauses such as data minimization, independent audits and risk assessments for school AI.
Operating questionWhich AI tools that process student or staff data are covered by an adopted policy and a contract clause, and which rely only on vendor promises?
Supporting evidenceSandy Valley Local School District (Ohio), as reported by Your Ohio NewsElectronic Privacy Information Center (EPIC), testimony to D.C. Council Committee of the Whole
Evidence ledger
Full record · every source keeps its link and limitations
Long Beach Unified surveys students and families on classroom AI and screens and forms a workgroup, including anti-tech parent groups, to shape policy
Long Beach Unified opened student and caregiver surveys in October asking how classroom technologies, including generative AI, affect learning, well-being, privacy and attention. A district spokesperson said a workgroup of teachers, students, guardians, administrators and industry experts, including members of grassroots groups lobbying to cut classroom tech, will review results after the survey closes on Nov. 6; findings, focus groups and research may shape policy recommendations. A Schools Beyond Screens chapter leader wants an AI moratorium like those in New York City and Los Angeles Unified. The district spent $2.5 million on laptops in 2025-26 and projects the same this year.
Why it matters, evidence and limitations
- Why it matters
- Directly K-12 and current: a large California district is moving from grassroots pressure to a formal input process that could produce AI restrictions. It follows the student anti-AI campaign archived on Oct 19 and adds the district's concrete response (surveys, workgroup, Nov. 6 close).
- Evidence and measured results
- News brief citing a district spokesperson and an advocacy leader; no survey instrument, response counts or draft policy were published in the article.
- Limitations and uncertainty
- Short brief; survey design, sample and timeline for board action are not reported. Moratorium references to NYC and LAUSD are the advocate's framing, not verified here. Survey responses are opinions, not outcome evidence.
Bloomington (MN) ISD 271 board approves small Colleague AI teacher pilot with a University of Washington study of student-learning effects
On Oct. 26 the ISD 271 school board approved a resolution authorizing administration to sign an agreement with Colleague AI Inc. and the AmplifyLearn.AI Center at the University of Washington to offer, implement and study the Colleague AI platform with a small group of teachers. The study is to examine whether platform use is tied to measurable gains in student learning. Colleague AI, a spin-off of a UW center funded by IES and NSF, markets assistants for lesson planning, assessment building, grading and differentiation. Schools, subjects, teacher count, cost and results timeline were not disclosed.
Why it matters, evidence and limitations
- Why it matters
- Directly K-12: a district is pairing a teacher-facing AI pilot with outcome research rather than adopting on vendor claims. The research partner is also the vendor's origin, which matters for independence; the archive holds the UW/Colleague AI Washington midyear report (+0.025 SD reading).
- Evidence and measured results
- Local news report of a board action; the district's Oct. 26 business meeting agenda (opened separately) lists 'Agreement with Colleague AI Inc. and the University of Washington to Implement and Study the Colleague AI Platform'. Contract terms and study design were not available.
- Limitations and uncertainty
- The outlet appears to be an automated or AI-assisted local news site (byline 'JR'); approval is reported there, while the agenda confirms only that the item was scheduled. Vendor marketing claims (thousands of educators) are unverified. No cost, scale or design details.
Sandy Valley (OH) school board adopts policy governing recording and AI processing of communications on school property and at school events
At its Oct. 19 meeting the Sandy Valley Board of Education adopted five new policies, including 'Recording and AI Processing of Communications on School Property and at School Events', which sets guidelines for audio and video recording devices, livestreaming and AI-enabled technologies. According to the report, it is intended to protect student and staff privacy while preserving legally protected recording rights and accessibility accommodations. Policy text was not included in the article.
Why it matters, evidence and limitations
- Why it matters
- K-12 and current: districts are starting to address AI note-taking, transcription and analysis of recorded conversations, not just student use of chatbots. This category (AI processing of communications at school) is rarely covered in district AI policies.
- Evidence and measured results
- Brief local report summarizing the policy's purpose from the superintendent; the policy text was not inspected.
- Limitations and uncertainty
- Policy text not available; the article summarizes purpose only. Ohio policy service model language may underlie it, which was not verified. Single small district.
EPIC testimony to D.C. Council urges school AI procurement guardrails: independent audits, risk assessments, data minimization and limits on surveillance mission creep
EPIC's Oct. 22 testimony for the D.C. Council oversight hearing on technology in schools argues that weak procurement lets harmful or unproven AI into schools and asks the Committee to legislate procurement standards. It recommends benchmarks on safety, privacy, security and equity before vendors may sell to schools, plus standard contract clauses: data minimization and short deletion periods, purpose limitation, district data ownership, cybersecurity and breach compliance, independent risk and impact assessments, and independent audits of training data, stress testing and output accuracy. It cites mission creep in AI monitoring, noting Lightspeed (a DCPS vendor) can notify law enforcement of possible self-harm, and accuracy problems in facial recognition and weapons detection.
Why it matters, evidence and limitations
- Why it matters
- K-12 and newly relevant: it arrived days before tonight's district items (Long Beach, Sandy Valley) and translates privacy concern into specific contract clauses districts can adopt. It is distinct from the archived CDT D.C. testimony of Oct 9.
- Evidence and measured results
- Advocacy testimony with footnoted citations (FTC orders, AP investigation, court settlements); it presents no new primary data. Claims about specific vendors rely on cited secondary sources not opened here.
- Limitations and uncertainty
- Advocacy document seeking legislation; not an evaluation. Dated Oct. 22, before the last run, and included because it is unarchived and directly relevant to current district procurement debates. Vendor-specific claims were not independently verified.
Frontiers meta-analysis of 58 GenAI learning studies finds large pooled effect but a prediction interval that includes zero and likely small-study inflation; only 7 primary-school effects
A three-level random-effects meta-analysis of 68 effect sizes from 58 experimental or quasi-experimental studies (6,941 participants, published 2023-2026) found a pooled g = 0.864 (95% CI 0.720-1.009) for GenAI on learning-related outcomes, with I² about 88% and a 95% prediction interval of -0.213 to 1.942, meaning a new setting could see null or negative effects. Egger's test showed funnel asymmetry and PET-PEESE suggested the true effect may be g ≈ 0.26-0.45. Most effects came from higher education (k = 50); secondary k = 11 (g = 0.612) and primary k = 7 (g = 1.311). Educational-level and outcome-domain moderators did not survive Holm correction.
Why it matters, evidence and limitations
- Why it matters
- K-12 relevant as a caution: headline effect sizes are often quoted to districts, but K-12 evidence is thin (18 of 68 effects) and the authors say local results could be null. Newly relevant because districts such as Bloomington are designing their own outcome studies this week.
- Evidence and measured results
- Peer-reviewed meta-analysis; English-language peer-reviewed studies only; many primary studies are small and language-focused; cultural-orientation contrasts are confounded with level and subject.
- Limitations and uncertainty
- Dominated by higher-education samples; K-12 subgroups are small; publication-bias corrections diverge (trim-and-fill vs PET-PEESE); between-study variance hit a boundary estimate. Published Oct. 21, before the last run; included because unarchived and newly relevant to district pilot design.
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.
- Academic research
- Research produced through an academic institution or peer-reviewed venue.