The public record of AI in scholarship VOL. 02 / 2026

The Public Record of AI Scholarship

AI writes the paper. A human expert judges it. The verdict — strong or weak — is published in full and kept permanently.

Open verdicts Permanent record Expert-reviewed Provenance logged
Generated by Claude
Civil Engineering Expert-reviewed

A Unified Framework for Predicting the Drag Coefficient of Natural Sediment Particles

We propose a unified model relating particle Reynolds number to drag across natural sediment geometries, reconciling empirical correlations from the prior literature into a single closed-form expression…

Expert verdict

“Follows the literature impressively — but the accuracy remains questionable.”

A. Riazi Promising · not yet rigorous
Auto-updating from the record
In plain terms
01 What it is

A public archive of AI scholarship

Every entry is a full-length paper written by an AI model — then judged, on the record, by a human expert in that field.

02 How it works

Idea to manuscript to verdict

Suggest a topic. One model is picked at random to write the paper. Experts score it. The verdict is published — and stays published.

03 Why trust it

Nothing is hidden

Weak verdicts sit beside strong ones. Each paper logs its model and knowledge cutoff. No cherry-picking, no quiet edits.

02 How it works

How the record gets made

From a research idea to a permanent, openly published evaluation — every step is transparent.

01 Human input
Your idea
Describe the article you'd like to see…
Any discipline Full manuscript

Submit an idea

Propose a topic in any discipline. We translate it into a rigorous, documented prompt protocol.

Suggest a topic →
→
02 Machine
Assigned · Claude
1 of 3

AI writes the paper

One model — Claude, GPT, or Gemini — is selected at random to write the full manuscript, its identity and knowledge cutoff logged as provenance.

Read the articles →
→
03 Human judgment
Expert review
✓Accuracy
✓Citations
✓Methodology
?Originality

Experts assess it

Domain experts score accuracy, citations, methodology, and originality — and publish the verdict openly.

Become a reviewer →
The outcome Idea → Manuscript → Review → A permanent public record
Why it matters

AI can produce a scholarly paper in minutes.

Whether it holds up is the question of the decade.

No benchmark can settle it — only an expert who reads the work and can tell the difference. So we're answering it in public, and keeping every receipt: each paper AI-written, each verdict signed by a human, nothing edited away.

The record, growing in public
81
Papers in the record
30+
Disciplines
3
AI models
03 From the corpus

What happens when AI tries real scholarship

Expert reviewers are building the only systematic record of AI scholarly capability. Here is what they are finding.

MIXED

Citation accuracy varies widely

AI produces impressively structured references, but verification reveals fabricated sources alongside legitimate ones.

Reliability Mixed
STRONG

Structure consistently strong

Abstracts, methodology sections, and logical flow rate well across models — the scaffolding of scholarship is convincing.

Consistency High
PENDING

Reproducibility remains unverified

Methods and calculations look plausible on paper. Whether they survive implementation is still an open question.

Verification Pending
In their words
Civil Engineering Claude

“The results clearly follow the same pattern observed in the literature, which is impressive; however, the accuracy remains somewhat questionable.”

Expert Reviewer · Amin Riazi
A Unified Framework for Predicting the Drag Coefficient of Natural Sediment Particles →
Social Sciences GPT

“I found the abstract very well structured: it provides contextual background, identifies the core issue, and clearly states what the article will offer.”

Expert Reviewer · Anonymous
Clarifying Research Quality Across Quantitative, Qualitative, and Mixed Methods →
Fundamental Sciences Gemini

“The manuscript looks very reasonable and the references are sound. But the validity of the conclusions is difficult to assess unless the calculations are reproduced.”

Expert Reviewer · Anonymous
Optimizing Sensitivity to Sub-GeV Dark Matter via Electron Recoil →
04 Why it needs you

It takes someone who can tell the difference

A model can imitate the shape of a paper flawlessly. Spotting where it quietly breaks — a fabricated citation, a method that won't reproduce, an argument that doesn't hold — takes a human who has done the real work in that field.

That judgment is the record. Every review you write becomes a permanent, citable line in the only systematic account of what AI can — and can't — do in scholarship. Whatever the answer turns out to be, there's a place in it for you.

Read our mission & vision →
Where you fit
R For researchers

Review AI-generated papers in your discipline. Your structured evaluation becomes a permanent, citable contribution to the only systematic record of AI scholarly capability.

E For educators

Assign students to review AI papers as a course exercise — they gain peer review experience while contributing to real research infrastructure.

∀ For everyone

Suggest a research idea in any discipline. Our system translates your concept into a structured prompt that generates a full manuscript for expert evaluation.

An open invitation

Your judgment belongs
in the record

Every expert who reviews an article writes a permanent, citable line into the only systematic account of what AI can and cannot do in scholarship. The record is only as good as the people who build it — and there's a place in it for you.

Reviewing alongside 40+ experts across 30+ disciplines