Guilford County Schools, a 120-school, 67,600-student district in North Carolina, launched the Guilford Guarantee in October, a promise that every student can earn a career credential, 12 hours of college credit, or a paid internship or apprenticeship before graduation, at no cost to their family, according to K-12 Dive. Superintendent Whitney Oakley told the outlet the initiative answers a specific number: roughly 70% of high school graduates nationwide leave with a diploma and no plan for what comes next, even as local employers struggle to hire. CTE credentials in the district are up 320% over five years, evidence Oakley points to as proof the old stigma attached to career-track courses has broken down.
Building it required two moves few districts attempt at once. Oakley and the president of Guilford Technical Community College committed to attending every planning meeting together for a year and a half, connecting CTE coursework directly to college-credit pathways that already existed but had never been mapped. The district also built employer pipelines from scratch: Toyota's new battery plant sits seven miles from a Guilford County high school, and the aviation company JetZero approached the district before it even had a local office, asking how to hire students directly into composite-materials work. Transportation, not curriculum, turned out to be the harder constraint, since college-credit access tracked closely with which students owned a car; the district is now adding transit stops with Greensboro and High Point's transit authorities to close that gap, while the nonprofit Shift_ed covers books and fees for students who can't.
Metro Nashville Public Schools, a district of roughly 80,000 students and 12,000 employees, partnered with Vanderbilt University's Data Science Institute to build PRISM, an AI system the district uses to audit its own non-instructional job structure, according to Education Week, reported via GovTech. PRISM runs five separate AI components that assess job titles and descriptions and flag inconsistencies, deliberately designed with what the district calls "architected friction," meaning the components are instructed to disagree with each other and force a second look rather than default to agreement. The system reviewed 992 distinct job titles covering more than 1,000 non-instructional positions and found, in one example, 435 employees sharing the title "specialist" spread across 24 different pay grades with inconsistent descriptions. PRISM's reclassifications matched human-level accuracy of roughly 80%, comparable to a staff analyst with two years of job-classification experience, while routing higher-profile or higher-paid roles to a person rather than reclassifying them automatically.
Wayne Birch, the district's strategic compensation and people analytics lead, framed the project as an HR fix, not an instructional one: "The focus on AI has been on student learning, and that's understandable, but my argument is that in order to meet those goals, you have to have quality people to both educate the students and support the educators." That distinction is the one worth borrowing. Most AI pitches landing in your inbox target the classroom directly; Nashville's most credible use case sits in the back office, finishing a task too large for a committee to do by hand while still sending the consequential decisions to a person.
A K-12 Dive feature on the long-running homework debate, reported by Ed Finkel and published June 10, 2026, gathered three researchers who converge on one conclusion: the right amount of homework is not a fixed number of minutes. Joyce Epstein, professor and co-director of the Center on School, Family, and Community Partnerships at Johns Hopkins University's School of Education, put it directly: "The point is not about the 'right number of minutes.'" Katie Newhouse, professor and director of special education programs at NYU Steinhardt, debunked the commonly cited "10 minutes per grade level" guideline as not research-based, recommending high schoolers receive 60 to 90 minutes of total homework a night, roughly 20 to 30 minutes per subject. Denise Pope, senior lecturer at Stanford's Graduate School of Education, added the sharper rule: homework should connect to a class's larger goals and should not double as an assessment. "You probably don't want to be grading homework," she said.
Newhouse's number, 20 to 30 minutes per subject, gives you something to check your own course load against tonight, not at a department meeting next semester. Pope's point cuts deeper: if you grade homework the way you grade a test, the assignment is doing a job it was never built for. Both researchers point toward flexible weekly deadlines and flipped-classroom structures as ways to account for students juggling after-school jobs, sibling care, or no quiet space at home, the inequities a fixed nightly deadline tends to punish hardest.
foundry10, a Seattle-based education research organization, and NORC at the University of Chicago released the first nationally representative survey of how American teens experience social media literacy education, surveying 1,027 high schoolers ages 14 to 18 between May and August 2025. Parents are both the most common and most trusted source of guidance, named by 76% and 69% of teens respectively; teachers followed at 57%, peers at 20%. Nearly three-quarters of teens had received instruction on risks like cyberbullying and misinformation, but fewer than six in ten had been taught about social media's actual uses or benefits. Jennifer Rubin, principal investigator of foundry10's Digital Technologies and Education Lab, said in a release that teens "who learned about both the risks and benefits of social media felt more confident and empowered online than those exposed only to fear-based messaging." Most teens also reported only partial understanding of how recommendation algorithms shape their feeds, and nearly two-thirds said the conversation should start before age 13.
The instinct to lead with warnings is reasonable, and the data doesn't contradict it; cyberbullying and misinformation are real risks worth naming directly. What the survey adds is the half of the lesson most curricula skip: the fewer than six in ten of your students who have ever been taught what social media is actually good for, or how the algorithm in front of them decides what they see. Rubin's framing is direct: "Adults have spent years debating whether social media is good or bad for teens, but far less attention has been paid to how young people actually learn to navigate these spaces. What we found is that many teens are not asking for fewer conversations about social media. They are asking for better ones."