Writing peer reviews with process evidence
Humanly Cloud can support peer-review writing through configured tasks and recorded drafting. Review organizers can provide instructions and resources, define the permitted in-platform AI assistance, and inspect submitted reviews with their process evidence. This helps explain how review text developed; it does not assess the scientific correctness or fairness of the review automatically.
What workflow is available?
A review organizer can create an assigned writing task and share it with a reviewer. The reviewer writes inside the configured workspace, then submits the response. The task owner can inspect the submission and its certificate evidence using the same underlying workflow as other assigned writing. See task setup and review for the shared controls.
For connected OpenReview assignments, Humanly also provides an OpenReview workflow. It uses the Humanly OpenReview Connector and a signed-in OpenReview identity to obtain accessible assignments and the venue's review form. Availability depends on that connection and the assignments available to the reviewer. This is not a promise of integration with every conference, journal system or review form.
A reviewer's sequence
- Confirm the venue's confidentiality and AI-use requirements before loading manuscript material into a writing or AI tool.
- Open the assigned Humanly task, or use the OpenReview connection flow in the Cloud application for an available connected assignment.
- Read the instructions, supplied resources and AI permissions. Use only assistance permitted for the review.
- Draft and check the review, including the required fields, ratings and factual references.
- For a connected review, use the explicit OpenReview sync flow and verify the resulting review on OpenReview. Check the final version again after later edits.
Saving a Humanly draft, generating a certificate and publishing a review to a venue are separate events. A local document or certificate alone is not confirmation that OpenReview received the current review. Humanly exposes sync status, but the venue's actual review record is the authoritative place to confirm external delivery.
How can a chair use the evidence?
A chair reviewing a configured task can compare the writing record with the task's stated permissions. If AI polishing was allowed, a recorded polishing action can be expected behavior. If a passage needs clarification, the available timeline and replay may help place that passage in its drafting context.
The evidence does not establish that the reviewer read the manuscript thoroughly, found every error or formed an independent scientific judgment. A typed passage can still reproduce outside material. Behavior signals and model scores need human interpretation, as explained in the certificate guide.
Confidentiality and sharing
A manuscript, draft review or AI prompt may contain confidential material. Check the venue's rules before choosing AI assistance or creating a publicly accessible certificate. The existence of a feature does not grant permission to use it for a particular venue.
Certificate display settings control whether the shared view includes full text and edit history, with an optional access code. These settings do not replace a venue's confidentiality requirements or control copies already shared. The Humanly Cloud privacy policy explains that AI features may process uploaded material, prompts and related content using service providers.
The research guide links the Humanly paper and distinguishes its studies from product guarantees. For the assistant's actual scope and the limits of observing external AI use, read AI policy.
Reviewed: 2026-09-24.
Applies to: Humanly Cloud release baseline 75eba744, checked against source revision d5818c71 on 2026-09-24. Product descriptions cover capabilities shared by that release and source, not a new deployment. Available controls and evidence depend on the writing setup. Research findings refer to arXiv:2607.21758v1.
Canonical: https://cloud.writehumanly.net/reference/peer-review