What exactly is being reviewed?
A specific work and version with a persistent Record identity.
AsMade for Reviewers · start with context, not a verdict
When AI is part of the work,what exactly are you evaluating?
The final work shows the outcome. It may not show who directed, what AI produced, what was checked or rejected, or which claims are supported.
A MADE Record gives you a structured view of one specific work, with deeper process, evidence and limits available when needed.
REVIEW LENSOne Work · layered context
Review without reconstructing everything first
Start with the main picture. Trace only what matters to your decision.
The Record separates the work itself, contribution, supporting material and known limits so a reviewer can move from overview to detail.
Identify the exact work first.
Each MADE Record is tied to one defined work and version rather than applying a person-level label or score.
Evidence Status and access state remain separate.The final evaluation stays with you.
The review problem
The same “AI used” label can hide very different processes.
One person may use AI only for wording. Another may use it for source images but make the creative direction, selection and assembly decisions. Another may use AI across research, drafting, translation and revision. A binary disclosure alone does not explain those differences.
Without a RecordYou may have to infer context from the final result or reconstruct it from chats, drafts, screenshots and explanations.
With a MADE RecordYou can start from an organised account of the Work, contribution, support and uncertainty, then inspect deeper material selectively.
Four questions before you decide.
The Record is designed to support judgment without making that judgment itself. It does not collapse different kinds of evidence or uncertainty into a single “verified” label.
Who did what?
Human actions, AI assistance, editing, selection and significant decisions can be distinguished.
What backs each important finding?
Statements can link to available Materials and Evidence Status while access remains separately controlled.
What is missing or uncertain?
Gaps, restricted material and unresolved questions remain visible instead of being silently interpreted as certainty.
Review the product, not a promise about it.
The two current public MADE Records show how the same reading model behaves across a creative work and a professional document.
The Count of Monte Cristo — Page 050
Direction, AI-assisted generation, rejection, selection and assembly are presented with differentiated support and an explicit limitation.
Professional CV
Source information, AI-assisted wording and translation, creator review and privacy-safe redacted evidence provide a different review context.
Reviewer boundary
More context. Still your judgment.
Not an AI detectorNot legal certificationNot proof of authorshipNot a truth guaranteeNot a quality or acceptability verdict
Help us test the AsMade MVP from the reviewer side. Apply for the pilot if you review AI-assisted work and a MADE Record could help. No work files are needed at this stage.
Apply for the MVP Pilot