Assess the thinking, not just the answer.
Students solve real problems and defend their decisions. AI asks the hard questions. Faculty make the final judgment.
A new skincare brand in Multan has Rs. 5,000 a month for social media. How should it be spent?
Why it matters
When AI writes the answer, the student's thinking disappears from view.
Distributed epistemic displacement
Students hand their reasoning to AI tools, so a polished submission no longer shows what they actually understand.
AOUNVA Assessment Studio reverses the role. The student does the thinking; the AI only asks questions.
How it works
Five stages, one connected line of thought.
- 1
Understand
Read a real scenario and take a first position.
- 2
Reason
Explain the evidence and assumptions.
- 3
Decide
Choose an action and name its trade-offs.
- 4
Defend
Answer the AI facilitator's challenges.
- 5
Reflect
State what could change the decision.
What we assess
A clear record of how each student thinks.
Every response is mapped to five dimensions of thinking. AI suggests a rating; faculty review the evidence and decide.
Reasoning
Are ideas logical and backed by evidence?
Strategic thinking
Are context, constraints and alternatives considered?
Decision quality
Is the choice justified, with trade-offs stated?
Response to challenge
Can the student defend or revise a position?
Epistemic awareness
Does the student recognise uncertainty and missing evidence?
AI asks. Students think. Faculty decide.
The AI
Facilitates
Asks follow-up questions and challenges assumptions. It never gives the answer.
Students
Think and defend
Work through a real problem, explain their choices and reflect on what they learned.
Try the student experienceFaculty
Hold authority
Review each exchange, add comments and set the final result.
Open the faculty dashboardBuilt on the AOUNVA toolkit
Where does the thinking go when AI can reason?
AOUNVA calls this Distributed Epistemic Displacement: knowledge and agency quietly moving from students to AI. The Studio keeps them with the student, using the methods of the AOUNVA AI-Resilient Assessment Toolkit.
Faculty score each assessment on ten questions before publishing and see which fixes lower the score.
Every assessment states Level 0, 1 or 2, so students know exactly which AI use is allowed and why.
Students commit to the rule before they begin and disclose any AI they used in the final check.
Faculty mark ownership, justification, response to challenge and awareness of limits.
Start with one course.
Run a focused pilot. Measure student experience, faculty workload and assessment quality before wider adoption.
Knowledge is that which liberates.