How this place works, in writing.

Every submission passes automated screening, then an advisory AI panel, then a subject editor, then at least two human reviewers with matching expertise. Only the editor can accept or reject.

Review is single blind by default: reviewers see author identities, authors do not see reviewer identities. Reviewers may choose to sign their report. Reviewers must declare conflicts of interest and are removed if one exists.

Authors receive a written decision letter with the reviewer reports attached. Where authors opt in, those reports are published alongside the accepted paper.

Decisions may be appealed. An appeal is handled by a different editor than the one who made the original decision, and you get a written response either way.

Five AI agents review each submission: screening, methodology, novelty and prior art, research integrity, and ethics. Their output is structured, stored, and labelled advisory.

An AI report cannot accept or reject a paper, and cannot overrule a human. If an editor disagrees with every AI report, the editor prevails and the disagreement is recorded.

Authors and reviewers see the model used, the prompt version, the full report, and the token cost. Nothing about the AI layer is hidden from the people it affects.

Submissions are never used as training data, by us or anyone else. Paper text is treated as untrusted input and sandboxed from tool access, because a submitted document can attempt to manipulate a model that reads it.

Authors must disclose AI assistance used in producing their own work. Using AI to help write is allowed; presenting AI-generated results as original findings is misconduct.

Every version is checked for textual similarity against published literature, AI-generated text signals, image manipulation and duplication, and duplicate submission by file hash.

Reference lists are verified for existence. Fabricated citations are treated as misconduct, not as an error.

Suspected misconduct is investigated by the integrity officer following COPE-aligned procedures. Authors are given an opportunity to respond before any finding is published.

Corrections and retractions are published as public notices explaining what changed and why. We do not delete records. A retracted paper keeps its DOI and is clearly marked as retracted, because the citation record must stay honest.

Anyone can report a concern about a published paper. Reports can be made in confidence.

The minimum author age is 15. We do not create accounts for anyone younger, with or without parental permission.

Where a jurisdiction requires parental consent below a given age — 16 in several EU member states and comparable rules elsewhere — we collect verified guardian consent before a submission enters review, and record the signature, timestamp and IP address.

For authors under 18 we apply stricter privacy defaults in line with the UK Children's Code: minimal data collection, no behavioural profiling, and no public display of anything beyond what publication requires.

Because a DOI is permanent, we offer a pseudonym option prominently at submission rather than burying it in settings. Guardians consent specifically to public attribution, not merely to account creation.

We collect what publication requires and little else: your name, email, date of birth, career stage, optional affiliation and optional ORCID iD.

Data is encrypted at rest and in transit. Access is limited to the smallest set of people who need it, and every administrative action is written to an append-only audit log.

You can export your data or request erasure at any time. Erasure removes your account and unpublished material. Published papers keep their DOI, since permanence is the point of a registered identifier; on request we will replace your name with your pseudonym.

We do not sell data, run behavioural advertising, or share personal data with third parties beyond the processors needed to run the service.

Submitting a paper grants EDA a non-exclusive licence to review, publish, archive and distribute it. You keep copyright. Accepted papers are published under CC BY 4.0 unless you select another Creative Commons licence at submission.

You must be the author of what you submit, and it must not be under review elsewhere.

We may suspend accounts for misconduct, abuse, or attempts to manipulate the review process, including attempts to influence AI reviewers through content embedded in a submission.

The service is provided as is. We aim for permanence and deposit to independent archives so the corpus survives us, but we cannot guarantee uninterrupted availability.