What changed on August 2, 2026
As of August 2, 2026, the EU AI Act officially classifies AI systems used for school admissions, grading, and exam monitoring as 'high risk.' That classification isn't just a label — under the EU AI Act's risk-tiered framework, 'high risk' systems come with concrete compliance obligations, including proof of human oversight and transparency.
In practice, that means schools using AI to help decide who gets admitted, how assignments get graded, or whether an exam is being monitored honestly can no longer simply deploy these tools and trust the output. They now need to be able to demonstrate that a real person is overseeing the AI-assisted decision, and be transparent about how that oversight works.
Why admissions, grading, and exam monitoring specifically
These three areas share something in common: they're all high-stakes decision points that directly affect a student's academic outcomes and opportunities. Regulators classifying them as 'high risk' reflects a broader pattern in AI governance — the systems facing the strictest oversight requirements tend to be the ones making or influencing consequential decisions about people's lives, not lower-stakes tools like a study aid or a homework helper.
What schools need to do
For schools operating in the EU, this requirement means:
Documenting who is responsible for reviewing AI-assisted admissions, grading, and exam-monitoring decisions
Building in an actual human review step, not just automated processing
Being able to show transparency about how these AI systems are used and supervised
The exact operational details are still being worked out school by school, but the direction is clear: AI can be part of these processes, but it cannot run unchecked.
What this means for parents and students
If your child's school uses AI to help grade work, monitor exams, or process admissions, there should now be a documented human check built into that process. It's a fair and reasonable question to ask your school directly how they're complying with this requirement, especially heading into a new school year.
The bottom line
The EU AI Act's 'high risk' classification for school admissions, grading, and exam-monitoring AI marks a concrete step in how education-sector AI is being regulated globally. Schools need to show real human oversight, not just automation — and families have a reasonable basis to ask how their school is meeting that bar.
▶ Watch the 30-second video version: https://youtube.com/shorts/lzXX7NuljZg
Want practical AI guidance for parents and educators every week?
Subscribe: https://www.aibyage.com/?modal=signup&utm_source=beehiiv&utm_medium=newsletter