A community policy by and for students, teachers, administrators, parents and caregivers
ARPS’s operative document on student technology use is policy IJNDB and its implementing regulation, I7, Appropriate Use of Technology, predates 1:1 devices, adaptive learning platforms and AI, and says nothing about screen time, developmental appropriateness, take-home devices or generative AI.
The Amherst School Committee’s Annual Goals for SY 26-27 list “Technology Policy (phones, screen time, AI, etc.)” under goal 5, Curriculum Review, alongside an I-Ready review, and commit the committee to “regular surveys and feedback loops” and “public forums.” The superintendent’s September 15 report announced a Techless Thursdays initiative and said it “will take time to plan and launch so that it supports instruction and reflects the needs of each school community.”
On November 3, 2026 the Amherst School Committee voted unanimously to establish Technology in Amherst Public Schools (TAPS), a policy subcommittee charged with advising the School Committee in the development of a new district technology policy addressing student screen use, digital learning tools, and artificial intelligence in schools. TAPS can include up to 15 representatives from school staff, district IT and instructional leadership, caregivers and families, students as appropriate, and School Committee members. Its deliverables are a guiding principles document and a policy recommendation to the School Committee, updated annually. Read the TAPS charge here. The Regional School Committee has not yet taken up the charge.
On August 10, 2026, Google made the Gemini tab in Google Classroom available to students of all ages and turned it on by default.
Google’s own researchers had warned about this two years earlier. In April 2024 more than fifty Google DeepMind researchers and ethicists wrote that conversational AI poses particular risks to children, who struggle to distinguish AI from human interaction, and could cause “diminished intellectual engagement” and a “reduced sense of personal competence” (Gabriel et al.). Google’s own learning team wrote the same year that general-purpose helpfulness “may often be at odds with pedagogy and learning,” because models give answers away and create “an illusion of mastery before it has been achieved” (Jurenka et al.). Two years later Google turned Gemini on by default for every grade. Common Sense Media rates the under-13 version of Gemini High Risk, describing it as an adult product with filters.
AI Overviews and AI Mode in Google Search cannot be disabled by an administrator; the only Google-provided option is blocking Google Search entirely.
A July 2026 Common Sense assessment of those features, run on accounts configured as 11- and 15-year-olds, found the AI completed every one of 360 homework assignments it was handed and rated the product an unacceptable risk for children.
On November 3, 2026 Google announced that Gemini and Google Search, including AI Overviews and AI Mode, will be turned off by default for K-12 student accounts beginning in early 2027, and that students in districts which keep AI off will lose access to Google Search entirely. Until then the August default stands. When the change arrives, every AI tool on an Amherst student account will be there because a district administrator turned it on.
Privacy and exposure risks are acute. Amherst parents report that young students have been exposed to pornography and other disturbing content on school devices. YouTube is available on Chromebooks. Canva AI image generation is available. Kids are finding workarounds to content blockers, accessing TikTok and who knows what else. Even if your student isn’t doing this, other students access this content on school devices and then show it to their friends.
We compiled a research brief with more than 180 sources. Read the full research brief here. These are the top-line takeaways:
Reading on screens produces a small but reliable comprehension deficit compared with paper. A 2018 meta-analysis of 54 studies and 171,055 readers found a paper advantage that the authors describe as roughly two-thirds of a year of elementary reading growth, larger for informational text and growing in more recent studies (Delgado et al.). A 2023 meta-analysis of 26 studies and 469,564 readers found the link between leisure digital reading and comprehension six to eight times weaker than for print, and negative at primary and middle school ages (Altamura et al.).
Expected tangible rewards, the coins, badges and streaks built into many classroom apps, undermine children’s intrinsic motivation, and the effect is larger for children than for adults (Deci, Koestner and Ryan, 128 experiments). Attractive but irrelevant details interfere with learning (Cheng et al., 2026). The APA’s 2026 expert report puts it plainly, that “features designed to capture attention can pull attention away from learning” and that engagement is not learning.
Using devices to calm children aged three to five predicted higher emotional reactivity three months later, strongest in boys and in temperamentally reactive children (Radesky et al., JAMA Pediatrics). A 2024 cohort found more tablet use predicted more anger a year later, and more anger predicted more tablet use the year after that. The AAP’s 2026 policy statement links solo, noneducational media use to poorer sleep, language and social-emotional outcomes, and engagement-driven design to prolonged use that displaces sleep, play and in-person interaction.
The strongest displacement evidence is on language. A meta-analysis of 42 studies and 18,905 children found quantity of screen use correlated negatively with language, while adult interaction around content correlated positively (Madigan et al.). Physical manipulatives beat symbol-only instruction in 55 studies, especially on retention (Carbonneau et al.). UNESCO’s formulation is that technology “should not substitute but instead complement face-to-face interaction with teachers.” Every minute on a screen displaces what the evidence proves works best for learning and development, hands-on learning, discussion and peer interaction.
No randomized trial shows a benefit from a generative AI chatbot for K-6 general education students. UNESCO’s 2023 global review found “little robust evidence on digital technology’s added value in education,” noted that “a lot of the evidence comes from those trying to sell it,” and set age 13 as the floor for generative AI in classrooms. The trials that do exist, on older students, point the other way; two of them are charted below.
The federal filtering law requires a filter to be installed and working. It sets no standard for how well it works. The vendors describe blocking as a reactive, whack-a-mole process. Google’s own documentation says its YouTube controls do not cover embedded videos, signed-out viewing or links pasted into a shared document. The methods to bypass filters are easily found on YouTube and TikTok.
Sources: Center for Democracy and Technology, “Out of Step” (January 2025; 1,316 students, 1,006 teachers, 1,028 parents); Common Sense Media, “Teens and Pornography” (January 2023; 1,358 teens). Both cover secondary students. No survey of elementary students or K-6 teachers on filter bypass exists anywhere; a teacher survey in Amherst would be the first.
Prepared from public documents. Nothing on this page is legal advice.
Massachusetts has no state guidance on elementary screen use, and the stalled cell phone bill explicitly excludes school-issued devices. DESE’s AI guidance leaves every decision to districts. Other Massachusetts districts (Cambridge, Arlington, Northampton, Acton-Boxborough) have moved on their own in 2026, and their moves are the nearest precedents.
The common ingredients of every 2026 policy are no 1:1 devices in the earliest grades, with shared carts instead; devices that stay at school; a daily cap or an “earn its way in” evidence standard for apps; no screens as rewards, calming tools or filler; explicit IEP and 504 carve-outs; parent notification and opt-out; and no student-facing generative AI in elementary. Iowa’s and Utah’s statutes and New York City’s policy are the cleanest model language to hand ARPS. San Francisco, which signed an AI contract before board review and put an AI reading tutor into K-5 without consulting teachers, has had a teachers’ strike and a thousand-signature parent petition, and its policy vote is not until March 2027.
| Policy element | New York CitySept 2026 | Los AngelesJune and Sept 2026 | Fairfax County, VAJuly 2026 | SeattleSept 2026 | Iowastatute, July 2026 | Utahstatute, July 2026 | Cambridge, MA2026 | Arlington, MA2026 | Acton-Boxborough, MAMay 2026 |
|---|
Cambridge’s School Committee ordered a usage audit after parent concerns, and the district already bars screens as a behavioral reward. Arlington moved to shared devices through grade 2 and keeps devices at school through grade 5. Acton-Boxborough adopted a phased AI plan in May under which elementary and junior high have no direct student AI tool use; this is the closest Massachusetts precedent for an ARPS AI policy. Northampton accommodated about a dozen opt-outs with analog alternatives after parents objected to YouTube replacing read-alouds. Melrose teachers made paper the default. Pioneer Valley Regional will remove laptops and tablets from preschool through grade 2 classrooms in January 2027.
The Amherst School Committee established the TAPS policy subcommittee on November 3, 2026. These are the working group’s recommendations.
Get the advocacy newsletter with meeting dates, what the committee decided, and what to do next.
Email TechnologyInAPS@gmail.comTell us what your child has encountered on a school device, or what you’ve seen in a classroom.
Email TechnologyInAPS@gmail.com. Responses are reported in aggregate and individual responses remain anonymous.Anything about the research, the policy proposal, the district’s current setup, or how the process works.
Email TechnologyInAPS@gmail.comHelp with research, outreach, the teacher survey, or speaking at public comment.
Email TechnologyInAPS@gmail.comThe Amherst School Committee takes public comment at every meeting, three minutes per speaker. We can help you prepare. How to make public comment, step by step.
Written comment: SCPublicComment@arps.org or voicemail 413-362-1891 by 3:00 PM on meeting day