Researchers led by Yossi Ben-Zion at Bar-Ilan University, working with colleagues at the University of Colorado Boulder and the University of Georgia, ran a controlled comparison in a college physics lab on electric potentials. One group used physical lab equipment. One group used a prebuilt simulator. One group prompted an AI tool to generate its own simulation of the phenomenon, then refined it through testing. On a conceptual assessment afterward, both the AI-generated and the prebuilt simulation groups scored significantly higher than the physical equipment group. The study, published in January in Physical Review Physics Education Research, reported a large effect size (η² = 0.359).
This is not an argument to throw out lab equipment. It is evidence that the value of a physics lab was never really the apparatus, it was the conceptual model students built while using it. A simulation, AI-generated or not, can build that model faster and without the setup time that eats into a fifty-minute period. The added benefit of the AI-generated condition was the process itself, students who had to specify what the simulation should do reported developing modeling skills the prebuilt simulator never required.
Alpha School, a private K-8 network that replaces classroom teachers with AI-driven software and human "guides," opened a Chicago campus this fall at the Lakeshore East building, charging $55,000 a year. Founder Mackenzie Price says the model lets students master core academics in two hours a day, freeing the rest of the day for life skills and projects. CBS Chicago reported in March that the school had drawn 35 interested families and enrolled two students toward a fall goal of fifty. Price cites internal data claiming students place in the top 1% nationally and grow 2.6 times faster than average on MAP testing, figures the company has not had independently verified.
The model's central claim, that AI can deliver core instruction in a fraction of the school day, is the same claim showing up in pilot programs and vendor pitches aimed at public districts, just without the $55,000 price tag or the guide-to-student ratio a boutique private school can afford. Liz Gerber, who directs Northwestern's Center for Human-Computer Interaction and Design, told CBS Chicago she is skeptical the approach develops the social and collaborative skills public school teachers spend years building. Two enrolled students against a fifty-student goal is itself a signal. The market is not yet convinced, and your district does not need to be either before vetting a vendor making similar promises at scale.
The Government Accountability Office reviewed five meta-analyses measuring the effectiveness of K-12 teacher professional development and found genuine disagreement in the research. Coaching, collaboration, training on curriculum materials, and pedagogical content knowledge were each linked to better student outcomes in at least one study, but not consistently across all five. Teachers themselves were clearer than the research: in a nationally representative RAND survey, 67% said collaborative learning opportunities, working and reflecting with other teachers, improved their teaching or their students' learning. GAO interviewed officials in three states and nine districts, all of whom described using Title II-A funding flexibility to meet local needs, and found districts are still more likely to run one-time training workshops than the sustained, job-embedded development the law was designed to fund.
You already knew the one-day in-service rarely changes what happens in your classroom on Monday morning. The GAO review gives that complaint research backing instead of just your own experience. It also names what does have evidence behind it: structured time to collaborate with colleagues, not another vendor demo. If your school's PD budget defaults to one-shot workshops, the GAO report and the RAND survey behind it are the citation to bring to whoever controls that budget.
Snapchat will no longer let users between 13 and 15 contribute videos to Spotlight, the app's publicly viewable short-form feed, or share Stories outside their confirmed friends. The company is replacing public access with a new "profile" feature where younger teens' content is visible only to mutual friends and strips out engagement metrics like favorite counts. Snap says the change is meant to encourage self-expression "within a trusted audience." Engadget reported the change on June 10. It follows several years of Snap adding parental controls and restricting how strangers can contact teen users, while the company continues to face lawsuits over its handling of child safety.
Your students under 16 just lost the audience-size incentive that made Spotlight worth posting to, the favorite count, the chance a stranger's video gets traction. Watch for two reactions in your classroom: relief from students who felt pressure to perform for a public feed, and workarounds from students who liked that audience and will look for ways to fake their age or move to a platform that still offers it. Neither reaction is hypothetical. Snap made this change because the open version of the product had a problem serious enough to fix.