Modern classroom with digital display wall

Grade the Human,
Not the Homework

What is Grade The Human, Not the Homework?

What happens when we stop grading the homework, and start grading student understanding instead?

Grade the Human, Not the Homework is a practical framework for scalably assessing student understanding, rather than solely grading an assignment submission. This approach moves beyond grading assignments to evaluating what students actually know.

Goals of a Human-Centered Assessment Strategy

A human-centered assessment strategy still demands independent practice, but grades students based on their understanding of it, not the submission itself.

  • Keep Authentic Practice

    Students complete independent practice and submit assignments as usual, maintaining the value of authentic learning experiences.

  • Grade Understanding

    Assessment focuses on student understanding rather than the submission itself, evaluating what students actually know.

  • Make AI Use Trivial

    AI use or non-use becomes trivial when we expect students to be accountable to their work, AI used or not.

Strategy Framework

Four assessment types for verifying student understanding at scale, with deployment options for every class size.

Interactive Check-ins

After independent practice and assignment submission, engage students in real-time conversations to assess their understanding through live discussions, spot-checks, or one-on-one sessions.

Literature

Spot Check

“Tell me about the main argument from the reading. Provide a counterargument.”

Business

Change a Variable

“What happens to your sales projection if interest rates increase?”

Humanities

Defend a Decision

“Why did you use the word 'marvelous' to describe the king's helmet?”

Browse Sample Assignments

Not sure where to start? Refine your preferences, then let us pick one for you. Or skip ahead and explore by discipline.

Not sure where to start?

Set class size, modality, and strategy type, then choose an example that matches.

Refine your pick

Filters apply only to Choose For Me.

Strategy type

Modality

Class size

Or explore by discipline

Choose a discipline to see one AI-vulnerable assignment and how all four strategies apply to it.

Literature and Writing

Start from one AI-vulnerable assignment, then see how each Grade the Human strategy grades understanding of that same work.

Typical AI-Vulnerable Assignment

Write a 5-7 page literary analysis essay on a course text. Develop a clear thesis, support it with textual evidence, and submit a polished final draft.

Apply All Four Strategies to This Assignment

  • Interactive Check-ins

    Talk About What You Know

    Ask the student:

    “Why did you prioritize this theme over the one on page 12?”

    Reveals whether the student owns their interpretive choices or only produced a polished surface argument.

  • Validation Checkpoints

    Modify or Troubleshoot Your Work

    Ask the student:

    “This claim-evidence link is weak. Revise it and explain what you changed.”

    A live repair of their own prose shows judgment about evidence, not just ability to generate clean paragraphs.

  • Process Demonstrations

    Show Me How You Did This

    Ask the student:

    “Highlight three key sentences and tell me why each one is essential to your argument.”

    Forces the student to surface the structural spine of the draft they built, not a summary AI could invent.

  • Reflective Defenses

    Explain the "Why" behind the "What"

    Ask the student:

    “Here is a competing thesis. Why does yours still hold, and what would you revise?”

    A rival reading exposes whether they understand their claim or only memorized the submitted essay.

Rubric for Grading the Human

The criteria describe the shape of an answer, not its content. Adapt the content standard to your discipline; the response-shape criteria can stay consistent.

Anchored

Is it inside their own work?

  • 3 · ConvincingLocates a specific sentence, data row, or region.
  • 2 · AdequateOne accurate, specific reference to their work.
  • 1 · ThinDiscusses the topic, locates nothing.
  • 0 · AbsentNo reference, or one that does not match.

Accurate

Is the content correct?

  • 3 · ConvincingContent and course terms correct throughout.
  • 2 · AdequateMinor slips that do not affect the claim.
  • 1 · ThinOne significant error, or a term misused.
  • 0 · AbsentFundamentally incorrect.

Reasoned

Is there a why, not a what?

  • 3 · ConvincingExplains the mechanism and what would change.
  • 2 · AdequateGives a reason but does not develop it.
  • 1 · ThinAsserts or describes, without reasoning.
  • 0 · AbsentNo reasoning offered.

Responsive

Did they answer the question?

  • 3 · ConvincingAnswers the question inside its constraint.
  • 2 · AdequateAnswers, but ignores the constraint.
  • 1 · ThinAnswers a nearby question instead.
  • 0 · AbsentDoes not engage the prompt.

AI Policy Language for These Assignments

Use this Demonstration-Based policy as suggested syllabus language for Grade the Human assessments. Copy it into your syllabus and adapt the bracketed details to your course.

Suggested starting language. Personalize for your course. Align any integrity language with your institution's academic integrity policies.

Demonstration-Based

When the grade comes from demonstrated understanding, so AI use polices itself.

Pairs with Grade the Human, Not the Homework

Suggested syllabus language

You may choose to use AI on this assignment, or not. That's up to you. Your grade will be based on a [follow-up conversation / oral defense / in-class checkpoint] where you explain, defend, and adapt your submitted work. Because credit comes from demonstrating your own understanding, using AI without learning from it will leave you unable to succeed. Use whatever tools help you genuinely learn the material.