← Back to Blog
🎓 AI in Education

AI in Education Digest — August 8, 2026

Six stories dominate AI-in-education news this week: a cheating scandal that just forced 158,000 university applicants to retake their entrance exam, America's biggest school district hitting pause on AI purchases, and new EU rules that force every classroom chatbot to say out loud that it's AI.

1. AI Cheating Scandal Forces 158,000 Students to Retake Mexico's Biggest University Exam

Source: CNN, Mexico News Daily, Times Higher Education · August 2–7, 2026

UNAM, Mexico's largest and most prestigious university, suspended undergraduate admissions for the 2026–27 academic cycle after its first-ever remote entrance exam produced a suspicious spike in near-perfect scores. On the medicine degree exam, the share of test-takers answering more than 100 of 120 questions correctly jumped to 29.3%, compared with under 10% in previous years.

An expert commission is now investigating suspected AI-assisted cheating during the online exam, which for the first time allowed candidates to sit the test remotely as long as their camera and microphone stayed on. All roughly 158,000 applicants have been told to retake the exam in person.

Why it matters: It's the starkest real-world proof yet that remote, AI-proctored testing is breaking down at scale, forcing universities everywhere to rethink online assessment.

If a camera and microphone can't stop this, what can?

2. Education's AI 'Gold Rush' Comes With Risks for Students, Christian Science Monitor Investigation Finds

Source: The Christian Science Monitor · August 7, 2026

A Monitor investigation puts the K-12 AI market at $730 million in 2026, projected by Future Market Insights to reach $18.5 billion by 2036, even as a growing body of research shows AI can erode children's ability to think for themselves. Common Sense Media separately rated the school version of Google's Gemini "high risk," finding it gives near-identical answers to a 6th grader and a high schooler and can surface inappropriate content.

The piece also notes that Los Angeles Unified's superintendent resigned after an FBI raid tied to an AI chatbot contract, and that Stanford's Generative AI for Education Hub has documented a "performance cliff" — students using unrestricted AI improve immediately, then their performance plummets once the tool is removed.

Why it matters: It shows the gap between marketing claims and independent safety testing is widening just as districts pour real money into unproven tools.

Follow the money, then follow the research — they're not pointing the same direction.

3. New York City Schools Pause AI Software Purchases Amid Two-Year Moratorium Push

Source: Chalkbeat New York, K-12 Dive · late July–August 2026

After 29 City Council members demanded a two-year AI moratorium and parents rallied outside a Panel for Educational Policy meeting on July 29, NYC Schools Chancellor Kamar Samuels told principals to pause new AI software purchases until the Department of Education's long-delayed AI playbook is finalized. Samuels has acknowledged the city "missed the mark" with its first guidance draft.

A broader petition seeking a two-year moratorium on generative AI in city schools has gathered thousands of signatures, and the final policy is expected to include stricter rules for younger students.

Why it matters: As the largest US school district, NYC's decision will likely set the template other districts follow — or resist.

When the biggest district in the country hits pause, everyone else is watching.

4. EU AI Act's Article 50 Transparency Rules Now in Force — and They Hit Edtech Chatbots Directly

Source: EdTech Innovation Hub · August 4, 2026 (rules took effect August 2)

New EU rules require any AI chatbot or agent that interacts with students or teachers to clearly disclose that it's AI — a statement buried in terms and conditions no longer counts. Providers must also add machine-readable marking to AI-generated content, and deployers must add visible labels to deepfakes. Fines run up to €15 million or 3% of global turnover.

Notices aimed at children must be child-friendly and age-appropriate, and a transition period for content-marking runs to December 2, 2026, while interactive disclosure duties have applied since August 2.

Why it matters: Every AI tool used in European classrooms — and any US edtech company selling into Europe — now has a hard legal disclosure bar to clear.

"I'm an AI" just became a legal requirement, not a nicety.

5. OpenAI Launches Three Role-Specific ChatGPT Plugins for K-12 Teachers, Professors, and Students

Source: EdTech Innovation Hub, The Register · August 4, 2026

OpenAI rolled out "K-12 Educator," "College Educator," and "College Student" plugins inside ChatGPT Edu and ChatGPT for Teachers, bundling pre-built workflows for lesson translation, exit tickets, syllabus design, and study-guide generation. The plugins connect to a user's own documents, calendars, and approved apps within institution-managed workspaces.

OpenAI says college-age users exploit only 1–10% of ChatGPT's capability compared with "power users," and that the plugins are meant to close that gap. The company is also the founding partner of a five-year initiative to train 400,000 US K-12 educators through the American Federation of Teachers.

Why it matters: It's OpenAI formalizing its push to become the default AI layer of the classroom, already deployed through partners like Houston ISD and Fairfax County.

Every major AI lab now wants a permanent seat in the classroom — this is OpenAI staking its claim.

6. Stanford's 2026 AI Index Shows Up to 84% of K-12 Students and 90–95% of College Students Now Use AI

Source: Stanford HAI AI Index, via EdTech Innovation Hub · August 2026

Stanford's flagship AI Index found student AI use has become the norm: between 50% and 84% of K-12 students now use AI for schoolwork, and adoption reaches roughly 90% in US higher education and 95% in the UK. Yet only 48% of institutions report having formal AI policies in place.

Among students and teachers actually working inside those policies, just 36% of students and 6% of teachers say the rules are clear. Meanwhile 52% of students say AI should be formally taught in schools, and up to 73% of high schoolers expect to use it as part of their learning.

Why it matters: It quantifies the exact gap driving every other story in this digest — adoption is outrunning governance.

Nearly everyone is using it. Almost nobody has written the rules.

🎬 Best story for today's video

158,000 Students Just Got Caught Using AI to Cheat — And It Changes Everything About Online Exams

The UNAM cheating scandal has the sharpest hook of the week: a concrete, dramatic story (158,000 students, a rigged-looking exam, a real institution retracting scores) that's inherently visual, easy to explain in 60 seconds, and taps directly into the "AI vs. exams" anxiety every student and parent is already Googling.

"A university let students take their entrance exam from home, camera on, AI everywhere — and almost 1 in 3 medical school applicants suddenly scored near-perfect. Here's what that tells us about the future of every test you'll ever take."

AI in education AI cheating UNAM scandal AI exam proctoring ChatGPT in schools EdTech news AI policy schools AI in universities academic integrity AI future of testing

Want tomorrow's AI-in-education digest straight to your inbox?

Subscribe to the newsletter

🚀 Put today's AI in education to work

Revising for GCSEs or A-Levels? RevisionLab turns your notes into AI-powered practice questions, flashcards and personalised study plans — built for the way you actually learn.

As an Amazon Associate I earn from qualifying purchases.