Final contribution report · worked example

Climate policy and economic inequality

Amara Ndiaye · Global Politics — Year 12

Construct and defend an evidence-based argument about the distributional effects of climate policy.

Mean human attribution

4.4/5

Provenance integrity

94%

Evidence segments

11

iWhat does the Contribution Metric measure?Expand

Con-Met measures the degree to which the human, the AI, and their interaction contributed to specific functions within the creation process.

Contribution scores are not grades and do not measure quality, intelligence, effort, correctness or academic achievement. They represent contribution attribution based on observable interaction evidence.

Con-Met measures contribution. The teacher evaluates learning.

Contribution dimensions

Human AI Emergent

Degree of contribution to each function, 0–5. Not a grade and not a quality judgment.

DimensionStudentAIEmergent
Understanding the task5/51/52/5
Question prompting5/52/54/5
Original ideas5/52/54/5
Research / input4/53/53/5
Critical evaluation5/52/54/5
Idea development4/54/55/5
Revision decisions5/52/53/5
Structure3/54/54/5
Language refinement3/55/53/5
Final judgment5/51/53/5

Student contribution attribution

The thesis, the challenge to the AI's strongest objection, the two external sources and the final formulation are all student-originated. The student rejected AI phrasing twice and narrowed an overstated claim without prompting.

AI contribution attribution

The partner supplied the initial counterargument, one reframing, a proposed section order and draft phrasing. It produced no submitted prose that survived unedited.

Emergent collaboration

The 'salience gap' — the distinction between measured and perceived incidence — did not exist in either the student's opening claim or the AI's first response. It emerged at minute 7 through the exchange and became the spine of the submission.

Questioning profile

Prompts moved from framing ('what is the strongest objection?') to assumption-testing and verification-seeking ('am I overstating this finding?'). No prompt requested generated content.

Critical evaluation profile

Three documented instances of challenging AI output; one instance of accepting a correction that weakened the student's own claim.

Idea provenance

Initial claim → Strongest objection → Challenge → Reframing → Conceptual leap → Refinement → Final formulation

Provenance integrity

Section 4 (design remedies, ~180 words) has no traceable development in the interaction record. Not an accusation — worth a two-minute conversation with the student.

Student reflection

“I started with a claim I thought was finished. The useful part was being argued with — the objection about salience is the reason the essay exists in its current form. I used the AI to stress-test and to phrase, not to think for me, and I can point to where each decision was mine.”

The two-step record

Contribution and quality, side by side

Step 1 · contribution metric

Objective attribution from the interaction record, applied identically to every student. Not a grade.

Recorded across 10 dimensions with 11 evidence segments.

Step 2 · academic quality

Judged by the teacher against “Global Politics — Argumentative essay rubric”.

Not yet assessed.

Open quality assessment →

The two results are reported together but never combined into a single number. Contribution does not raise or lower the grade, and the grade does not reinterpret contribution.

Metric method · expand
Contribution framework
Con-Met Standard
Evidence source
Recorded interaction history
Assessment type
Contribution attribution
Quality judgment
Not included
Teacher judgment
Separate

Objective, consistent, reproducible, auditable and neutral: the same interaction evidence and the same contribution rules produce substantially the same attribution, and every value links back to the evidence it came from.

Academic assessment · teacher-owned

Not produced by Con-Met

Contribution attribution is evidence that can inform assessment. It is never converted into a grade. These judgments remain with the teacher:

  • Content knowledge
  • Accuracy
  • Argument quality
  • Evidence quality
  • Creativity
  • Communication
  • Subject-specific criteria
  • Learning objectives
  • Final grade

Con-Met AI · Make thinking visible

Measure contribution. Trace provenance. Support judgment.

Con-Met does not determine how good a contribution was. It shows where contribution came from and how it developed.

The system measures contribution. The teacher evaluates learning.