Edition 19 | May 18, 2026
This week's theme: Trust, governance, and the hard questions schools are now forced to answer. Princeton ended a 133-year tradition because of AI cheating. A Palo Alto family is suing a school district over a faulty AI detector. Google caught hackers using AI to attack real systems. The Senate moved a bipartisan bill to ban AI companions for kids. Michigan published statewide classroom AI guidance. A Washington school district saved a quarter million dollars by building its own AI tools. Microsoft made an AI tutor free for every K-12 student over 13. Anthropic and the Gates Foundation committed $200 million to AI in global education. And in Orange County, parents packed a workshop to weigh in on their district's AI policy. The week is a snapshot of a system trying to catch up.
Princeton's faculty voted nearly unanimously on May 11 to require proctors at all in-class exams starting July 1, ending an Honor Code tradition that began in 1893. The reversal came after the dean's office reported a sharp uptick in cheating concerns over the last six months, with senior surveys showing 29.9 percent of students admitted to cheating and faculty linking the surge directly to generative AI making cheating harder for peers to observe and report.
Why it matters: A 133-year tradition does not change without significant pressure. K-12 districts can use Princeton's reasoning as a planning prompt: when peer reporting is the only enforcement mechanism, undetectable tools change the math. The most resilient academic integrity policies pair multiple forms of evidence with clear student-facing appeal pathways, and Princeton's pivot is a useful conversation starter for cabinet, faculty, and student government to land on a process every party can stand behind.
Google's Threat Intelligence Group announced on May 11 that it disrupted a cybercrime group's attempt to use AI to exploit a previously unknown vulnerability in a widely-used systems administration tool. John Hultquist, the unit's chief analyst, said the moment had arrived that cybersecurity experts had been warning about for years: malicious hackers arming themselves with AI to supercharge their ability to break into the world's computers. The tool exploited resembles capabilities from Anthropic's recently announced Mythos model.
Why it matters: Schools and districts are among the most frequent ransomware targets, and AI-assisted exploits raise the floor of what a small criminal group can do. Districts that have already built a written incident response plan, run a tabletop exercise this calendar year, and clarified who calls whom in the first 24 hours will recover faster when, not if, it happens. The best time to drill the plan is when nothing is on fire.
On May 14, Anthropic and the Gates Foundation announced a four-year, $200 million partnership combining grants, Claude credits, and technical support across global health, education, and economic mobility. The education portion is the largest single education commitment Anthropic has made: Claude-powered evidence-based tutoring for K-12 students, career guidance for students entering the workforce, and AI-supported foundational literacy and numeracy programs in sub-Saharan Africa and India.
Why it matters: When a foundation as influential as Gates anchors a partnership at this scale around evidence-based tutoring, the field gets a strong signal about which AI use cases are most defensible right now. Tutoring with verified curriculum and tight guardrails is the use case with the clearest research base, and districts evaluating tutoring vendors this year have a real opportunity to ask sharper questions: what evidence supports your impact claims, what guardrails are in place, and how is your roadmap aligning with these emerging philanthropic standards.
The Michigan Department of Education released formal guidance on May 12 helping local school districts use artificial intelligence in classrooms responsibly. Michigan joins a broader wave of state-level action: MultiState now tracks 134 AI-in-education bills in 31 states, with student data privacy, human oversight, and graduation-readiness as the three dominant themes. Illinois districts face a July 1, 2026 deadline to have an AI plan in place.
Why it matters: State-level frameworks set the floor for what districts must address in their own policies. Michigan's guidance, paired with the Illinois deadline pressure and the MultiState tracker activity, is a useful prompt for any district to pull its current AUP, the state guidance document, and any pending state legislation into one room this month. The goal is to land on a coherent district-level position that you can articulate to your board and your families with confidence.
Peninsula School District CIO Kris Hagel reported a projected $250,000 in savings from canceled edtech contracts after non-coder administrators built internal HR, finance, and workflow tools using Claude Code subscriptions at $200 a month per user. The district is on track to cancel three to four major SaaS subscriptions by the 2026-27 school year, including its current workflow automation tool. The practice, sometimes called vibe coding, lets staff describe what they want in plain English while AI writes the actual code.
Why it matters: Peninsula's experience is one of the clearest cost-savings stories in K-12 right now, and the practice is replicable for any district with a willing tech leader and a few hundred dollars a month. The opportunity is real, and the discipline matters: districts that pair vibe-coded tools with security review, data-privacy posture, and a written backup-vendor plan will capture the upside without giving up the safeguards. Worth a conversation with your CIO this month.
On May 11, a Palo Alto family filed a federal civil rights lawsuit after their high school sophomore was accused of AI cheating based on detection-tool output the family says was a false positive. The complaint asks for the student's final grade to be restored from a lowered grade back to the earned B, and for any reference or implication of academic dishonesty to be removed from his school record. The case is part of a national pattern: Vanderbilt has already paused its AI detector and other universities are under similar pressure as evidence mounts that detection tools disproportionately flag non-native English speakers and neurodivergent students.
Why it matters: AI-detection lawsuits are reshaping what an academic dishonesty process must look like. Districts using detection tools as standalone evidence are exposed; districts that pair tool output with conferencing, writing portfolios, and a clear appeal pathway are positioned to defend their decisions. The Palo Alto case is also a reminder that families now expect transparency about which AI tools are being used to evaluate their child's work.
The Senate Judiciary Committee unanimously approved the Guidelines for User Age-verification and Responsible Dialogue (GUARD) Act this week, which would prohibit AI companion apps from interacting with children and teens nationally. The bill explicitly exempts AI chatbots used for educational purposes, mirroring laws already enacted in Idaho, Oregon, Washington, and California. Federal preemption would set a national floor while leaving room for state-level education-specific carve-outs.
Why it matters: The education carve-out matters more than the headline. If GUARD passes, districts need to be able to demonstrate the distinction between an educational AI use and a companion app. Pull your AUP forward and ask: does the policy already define which AI tools are educational, which are not, and who has the authority to make that call. The clearer the boundary you draw today, the easier the compliance conversation will be later.
Microsoft made its Study and Learn Agent generally available inside Microsoft 365 Copilot at no additional cost for every K-12 and higher-ed customer, for students ages 13 and older. The agent turns Copilot Chat into a Socratic tutor: a student stuck on calculus is walked through the problem step by step without being given the answer, and a student studying for a biology exam can quiz themselves on the cell cycle with flashcards generated on the spot.
Why it matters: Free AI tutoring embedded inside Microsoft 365 is a real expansion of what schools already have available without adding a line to procurement. Districts already on Microsoft 365 should inventory whether Study and Learn matches a current tutoring or homework-help vendor they pay for separately, and whether the agent fits inside the FERPA and parental-notice posture already in place. Free is only free if it is also intentional.
On May 13, Anthropic released Claude for Small Business, a packaged set of 15 agentic workflows and 15 reusable AI skills built to integrate with QuickBooks, Canva, Google Workspace, Microsoft 365, and Slack. Anthropic also released a free on-demand AI Fluency course co-developed with PayPal, and is running a 10-city free workshop tour through the end of June, with stops including Chicago, Atlanta, Dallas, and Los Angeles.
Why it matters: Educators running side businesses, after-school programs, and consulting practices are a growing audience for small-business AI tools. The free AI Fluency course is also a useful reference document for school leaders writing AI literacy curricula, because it presents a clean 4D framework that maps directly to classroom-friendly conversations about discerning, deciding, delegating, and documenting AI use.
Orange County Public Schools held a public workshop on May 13 where parents weighed in on the district's proposed AI policy, with officials aiming to have an approved policy in place before the 2026-27 school year. The OCPS conversation mirrors a broader national pattern: districts are racing to publish formal AI policies, and parents are showing up to ask sharper questions than they have in prior tech-adoption cycles.
Why it matters: Districts that bring parents into AI policy conversations early, rather than presenting the finished policy at a board meeting, build durable trust that pays off when the inevitable incident arrives. The format OCPS used, a structured workshop with open feedback, is replicable for any district. Even one community session before the start of the year creates a real channel for parent voice and gives leaders a clearer picture of where the concerns actually live.
Try This Week
Run a 30-minute "AI policy stress test" with your leadership team. Pick three real situations from this week's headlines: a student claims a detector falsely flagged their essay, a teacher wants to start using Microsoft Study and Learn with seventh graders, and a parent emails asking which AI tools their child is being asked to use. Walk through each one out loud. Who handles the first call. What evidence does the district need. What does the family hear in the first 24 hours, the first week, and the first month. The Princeton, Palo Alto, Microsoft, and OCPS stories in this newsletter all point in the same direction: the AI conversation is no longer abstract. The districts that have already practiced these conversations will navigate them best.
Until next time,
Dr. Janette Camacho
CEO, iTeachAI Academy
P.S. iTeachAI Academy now has classes in all 50 states, each with a built-in AI guide that lets educators question the curriculum, pull classroom-ready scenarios for any lesson, and get the concept explained a new way until it clicks. State catalogs at classes.iteachai.co.
Free AI courses at classes.iteachai.co
17 free AI tools at iteachai.co/TeacherTools
Know a teacher who needs this? Forward this email.
Subscribe free at iteachaibot.com