Teachers in secondary education spend an average of 6 to 8 hours per week grading. In vocational and higher education, this increases to 14-15 hours. What if you could drastically reduce that time without compromising on quality? That’s what AI-assisted grading is all about.
AI as assistant, not as replacement
The principle is simple: AI analyzes answers and suggests scores and feedback. As a teacher, you review those suggestions and make the decision. This works for all question types – from multiple choice to essays – and for all subjects. Whether you’re assessing a physics formula, reviewing a historical analysis, or reading a French essay: the AI adapts to the level, the subject matter, and of course the scoring criteria.
The teacher in control
The EU AI Act sets clear requirements for AI systems in education: human oversight must always remain in assessments. No black box arbitrarily assigning grades. Examplary has been built from the ground up with these safeguards in mind. Five elements ensure you always maintain control:
- Scoring criteria and facts from source material
The AI doesn’t work in a vacuum. There are transparent scoring criteria and the AI extracts relevant facts from the teaching materials and sources. This allows the AI not only to say “this answer scores 2 out of 3 points,” but also to indicate which elements are missing or well-developed. - Transparent reasoning
For each criterion, you see why the AI suggests a particular score, including the AI’s “confidence level.” Which parts of the answer were considered? What’s correct and what isn’t? Is the answer right but using different reasoning or wording? This transparency allows you to evaluate the AI suggestion rather than blindly trust it. - Mandatory human oversight
This is not a ‘check-all-boxes’ system. For all answers that cannot be assessed with absolute certainty – including open-ended questions, essays, arguments – you must explicitly approve or modify the suggestion. The AI doesn’t make decisions, you do. - Horizontal grading
You can also grade per question instead of per student, completely anonymously. You then assess all answers to question 3, then all answers to question 4, and so on. This prevents you from being (unconsciously) influenced by previous scores or your expectations of a particular student. It also makes grading more efficient: you get into the rhythm faster of what constitutes a good answer to that specific question. - Partial points and nuance
No all-or-nothing scores. The AI can assess with nuance and indicate that a scoring criterion is only partially met, or that certain aspects are well-developed while others are missing.
Three ways to use it
📝 For summative assessments, the AI suggests grades and feedback. You critically review each suggestion and determine the final grade. The AI saves you time in formulating feedback and consistently applying criteria.
📊 For formative assessments, you quickly gain insight into where students stand. The AI suggests scores and feedback, allowing you to see at a glance which learning objectives need attention. No hours of grading before knowing how the class is performing.
💪 For practice tests, students receive direct AI feedback without your involvement. They can practice independently, you don’t need to grade. More practice opportunities without extra work.
Privacy and compliance built-in
In addition to control safeguards, Examplary also protects your privacy and that of your students: data stays in Europe and is never used for training. GDPR-compliant from day one.
More than just Examplary tests
The grading module also works for tests and assignments you’ve created outside Examplary. Through integration with your LMS, you can seamlessly switch between where students submit their work and where you assess it.
In practice now
This isn’t future talk. Examplary is currently researching AI-assisted grading together with De Goudse Waarden in Gouda, where teachers and students gain experience in an Innovation Lab. Similar pilots in vocational and higher education will start soon.
Want to experience what AI-assisted grading feels like? Create a free account or contact us at hi@examplary.ai to participate in a pilot.