AI in Education Digest
22 May 2026

In what may be the most significant EdTech announcement of 2026, Google partnered with ISTE+ASCD to launch the Google AI Educator Series — free AI literacy training for every single K–12 and higher education teacher in the United States. That's six million educators getting access to Gemini and NotebookLM training at no cost.

The series went live on May 13 with new modules dropping monthly. The training is deliberately "snackable" — sessions run 10–15 minutes for K–12 teachers and 30–45 minutes for higher ed faculty, broken into micro-trainings designed to fit inside a lunch break or prep period. Complete a session, earn a badge. It covers foundational AI understanding, pedagogical applications, and admin use cases — the full picture.

This is Google planting its flag deep in the education sector. If 6 million teachers start building lessons around Gemini and NotebookLM, that's a generational lock-in. And for students? The teachers who shape their learning are now going to be AI-literate whether they signed up for that or not.

Boston Public Schools made national headlines when Mayor Michelle Wu, Superintendent Mary Skipper, and tech entrepreneur Paul English announced a $1 million public-private partnership to make BPS the first major-city school district in America to guarantee AI proficiency for every high school graduate.

The programme, launching across 20 high schools before expanding district-wide, is not just a coding elective or a one-off assembly. It includes teacher training, an industry-informed curriculum developed with UMass Boston, student hackathons, internships, and real career pathways. English — co-founder of Kayak and a BPS graduate himself — is bankrolling the seed grant personally.

The phased rollout begins next academic year, with full district-wide implementation the year after. And with rollout planned for September 2026, this story is very much still developing.

The OECD's flagship Digital Education Outlook 2026 dropped a finding that's going to make a lot of teachers say "told you so." While students using general-purpose GenAI tools produce higher-quality outputs on assignments, those advantages tend to disappear — or even reverse — when AI access is removed, such as in exams.

Completing a task with AI assistance does not necessarily equal learning, the OECD is warning. Meanwhile, 37% of lower secondary teachers globally already use AI in their work, and 57% say it helps them write or improve lesson plans. Yet 72% of those same teachers are worried about academic integrity — specifically, students presenting AI-generated work as their own.

The report is careful not to condemn AI in education. Instead it frames the issue as one of design: when GenAI is used purposefully, with clear teaching principles, it can genuinely support learning. The problem is most schools are still in experimentation mode, not purposeful deployment mode.

The Council of the EU formally approved conclusions this month calling for an "ethical, safe, and human-centred approach" to AI in European education systems. The core message: AI should strengthen teachers, not replace them.

The Council's document calls specifically for strengthening digital skills and AI literacy across all ages, guaranteeing inclusion and fairness (so AI doesn't become a tool that widens the gap between well-resourced and under-resourced schools), and protecting teacher and learner well-being. It's a direct counterweight to the pressure many European education ministries are under to adopt AI tools quickly without proper frameworks.

This comes as European universities are simultaneously facing the EU AI Act's August 2026 deadline — which classifies AI used for student assessment, admissions screening, and academic progress monitoring as high-risk, requiring transparency, human oversight, and documentation. Many institutions reportedly are not ready.

Google published results from two new Gemini impact studies conducted in Sierra Leone and Italy — and the numbers are striking. In classrooms where teachers used Gemini for lesson planning and personalised instruction, 80–99% of students in each class successfully mastered their planned lesson skills, which ranged from geometry and parabola calculations to writing Java code.

Even more interesting: teachers reported a 70% reduction in time spent on administrative tasks. That's not time they banked — they directly reallocated it to one-to-one student mentorship, motivational support, and emotional check-ins. The AI handled the paperwork. The humans did the human stuff.

It's worth noting these are Google-funded studies about a Google product, so the usual caveats about independent replication apply. But the direction of travel — AI handling admin burden so teachers can focus on relationships — aligns with what teachers themselves say they want from EdTech.

Parents at a New York City Panel for Educational Policy meeting publicly called on the Department of Education to pause all AI deployments in schools, arguing that rolling out tools before the DOE's comprehensive AI playbook (due June 2026) is finalised puts students at risk.

The DOE's draft guidance already states that AI cannot be used to assign grades, make disciplinary decisions, or collect biometric and behavioural data without strict oversight, and that every AI tool must be vetted for bias and equity impact. But parents argue the process is moving backwards — tools first, guardrails second.

The controversy highlights a tension playing out in school districts everywhere: the pressure to not fall behind on AI adoption versus the responsibility to protect students until proper governance structures exist.

A widely discussed survey finding from early 2026 is getting fresh attention: 65% of educators are now using AI to bridge resource gaps in their classrooms — even as many report being exhausted by the constant churn of new platforms and tools. Of those using AI, nearly half (48%) use it for both student learning and administrative tasks.

The most successful classroom AI deployments are following what researchers are calling the "teacher co-pilot" model: AI acts as a real-time assistant generating differentiated lesson prompts and suggesting resources, while the teacher edits, approves, and delivers with human context and relationship. AI as co-pilot, not autopilot.

The survey also flags a structural problem: lack of system integration. Teachers are juggling multiple AI tools that don't talk to each other, creating a new kind of administrative burden rather than reducing the old one.