Sean Campbell Proposes Judging AI-Assisted Work by What People Can Defend

A polished essay, spreadsheet, report, or presentation may no longer reveal who did the thinking behind it. In a newly published argument, Sean Campbell proposes the Three Humans Framework, which would judge AI-assisted work by whether a person can explain and defend the decisions behind the final artifact.

The artifact is not the test

Campbell’s starting point is permissive rather than restrictive. He says students in his classes can use AI extensively to create essays, reports, spreadsheets, and presentations. His proposed answer is not to determine whether AI touched the work, but whether the student exercised judgment over it.

That distinction separates two outputs that may look identical on a screen: AI-generated work and work made with AI under human direction. Campbell argues that the deciding question is whether a person can account for the choices that shaped the result.

Three audiences, one explanation

The framework gets its name from three possible evaluators. One is the instructor, sometimes acting as a stand-in for a future manager. Another is an outside stakeholder with a genuine interest in the problem. The third is a fellow student who must understand the reasoning well enough to challenge it and collaborate on a team.

What the defense is meant to reveal

• Why the person made particular choices.
• Whether the assumptions can withstand scrutiny.
• Whether the person can identify weaknesses and answer a challenge.
• Whether new evidence can change the conclusion.

A more demanding kind of proof

Campbell’s proposed test is an oral defense, not a second written assignment. He argues that explaining choices, responding to criticism, and changing a view when evidence warrants it can show whether someone understands AI-assisted work.

The approach also puts a visible burden on the person submitting the work. Campbell acknowledges that public speaking, disagreement, and being wrong in front of others can be uncomfortable. But he argues that defending a recommendation and responding to criticism will matter more as AI improves at producing the work product itself.

The unresolved question is whether schools and workplaces will embrace a more conversational and demanding assessment. Campbell does not argue for removing AI from the process. He argues that the human contribution should remain legible when another person asks why.