The journal · Peer review
Peer review at the Commons
Every paper in the Journal of the Commons passes through one unusual pipeline: scrutinised inside the Commons that made it, screened by a panel of multiple independent AI agents, and judged by external peers — then opened to the world only once its result reaches the field.
Two houses
Commons, then the world
Every manuscript is judged first inside the Commons that made it, and then by independent experts outside it. Nothing is published on the strength of a single room's opinion.
Machine review
Many lenses, many machines
Before any human external round, multiple independent AI agents review every manuscript in parallel — each owning a lens, each returning a structured, auditable verdict the editor can read line by line.
Earned openness
Closed until it is true
A peer-reviewed paper stays closed to the Commons while its result matures, and opens to the entire world the instant that result is translated into the field.
The lifecycle
From submission to open access
- 01
Submission
Author · the Commons
A manuscript enters from within the Aitele Research Commons — there is no external submission portal. Work is lodged and carried through the Aitele dashboard as one of two kinds: a breakthrough, or an implementation that carries one into a use case.
- 02
Scrutiny by the Commons
Editorial desk
The editorial desk examines the manuscript for scope, originality and basic soundness — a desk gate asking a single question: is this a genuine contribution worth putting before reviewers? Work that does not clear scrutiny is returned to its authors with notes.
- 03
Accepted as preprint
PreprintOn clearing scrutiny the manuscript is posted as a citable, timestamped preprint. Its title, authors and abstract are public to the world; the full text follows the journal's access rules.
- 04
Pre-submission review by the Commons
Fellows & Patrons
Round one. The Fellows and Patrons of the Commons read and challenge the preprint in open, collegial review — the people closest to the field stress-testing the claims before the manuscript ever leaves the house.
- 05
Accepted as pre-submission print
Pre-submissionHaving satisfied the Commons round, the manuscript is advanced to a pre-submission print: a preprint that has survived internal peer scrutiny and is ready for formal review.
- 06
Review by a panel of independent AI agents
The AI panel
The defining stage. Multiple independent AI reviewer agents evaluate the manuscript in parallel — each in its own context, each owning a distinct lens, from novelty and method to statistics, reproducibility and the match between claims and evidence. Every agent returns a structured, auditable verdict; a manuscript must clear the panel's threshold to proceed, and the editor adjudicates dissent.
- 07
Proceed to external review
Independent experts
Manuscripts that clear the panel proceed to independent external peer reviewers — domain experts from outside the Commons — for the final, human round of scrutiny.
- 08
Accepted as peer-reviewed publication
Peer-reviewedOn passing external review the article is published as peer-reviewed, carrying a note that its translation to the field is still pending. At this stage it is read in full by the Fellows and Patrons of the Commons.
- 09
Translation to the field
The Commons
The result is carried out of the page and into a working use case — a deployed system, a fielded instrument, an implementation that puts the invention to work.
- 10
Made open access
Open accessThe moment a result reaches the field, its paper opens to the world to read — peer-reviewed, under a distribution licence appropriate to the result: permissive (CC BY) where the work is unencumbered, read-only where it has been licensed. Closed while it matures; open the instant it proves itself.
Stage 06, in detail
The AI panel
What sets the Commons apart is a review layer no committee could sustain by hand: multiple independent AI agents that read every pre-submission print in parallel, each interrogating the manuscript through a single, fixed lens — the lenses below.
Novelty & prior art
Has this been done before — and is the advance real?
Methodological soundness
Do the methods actually support the design and the claims?
Reproducibility
Could an independent group reproduce this from what is given?
Statistical validity
Are the statistics correct, sufficient and honestly reported?
Claims vs. evidence
Does every claim trace to evidence inside the paper?
Internal consistency
Do the abstract, results and conclusions agree with one another?
Related work & citations
Is the literature represented fairly and completely?
Limitations & validity threats
What could make these results wrong, and is it acknowledged?
Data provenance & ethics
Where did the data come from, and was it obtained properly?
Clarity & exposition
Can a competent reader follow and check the argument?
Independent
Each agent runs in its own context and never sees the others' verdicts — many independent reads, not one opinion rehearsed under different names.
Structured & auditable
Every verdict is a written record against a fixed rubric — a pass, a flag or a fail, with reasons the editor and authors can read.
A screen, not a judge
The panel decides what reaches human reviewers, never what is published. External peers and the editor make the call.
The panel is a screen, not an oracle. It cannot accept a paper, and it cannot reject one on its own authority — its verdicts decide only what earns the time of human external reviewers, and every flag it raises is recorded for the editor and the authors to weigh.
Read the journal, or bring it work
Full text is free to Aitele Fellows and Patrons, and to everyone for papers opened on translation. Submissions are accepted only from within the Aitele Research Commons.