AI4Meta Beta Notice

Effective date: 2026-06-28
Document status: beta-program disclosure for invited users

AI4Meta is professional research software for evidence synthesis, meta-analysis, systematic review, screening, extraction, coding, and research workflow support. This notice explains what the beta program permits, what it does not permit, how data may be processed, and what responsibilities remain with the researcher or research team. Please read it before accepting an invitation or using AI-assisted features.

This notice is not a substitute for a full Privacy Policy, Terms of Service, data processing agreement, institutional review board approval, ethics approval, or legal advice. If your institution requires one of those instruments before using external software, obtain it before uploading project materials.

Purpose and beta status

AI4Meta is in beta. The software is being actively developed, evaluated, and improved with invited researchers. During beta:

• Features may change, be renamed, limited, or removed.
• Extraction schemas, ranking logic, AI prompts, and workflow defaults may be revised.
• Re-analysis or re-extraction may be required when project schemas change.
• Performance and availability may vary while server capacity and research demand are balanced.
• Output should be considered preliminary until verified by qualified human reviewers.

AI4Meta is designed to support research work, not to replace scientific judgment, domain expertise, statistical review, or authorship responsibility.

Invitation, access, and availability

AI4Meta beta access is invitation-based. Applicants may submit a request through the public invitation request page or a QR code distributed by AI4Meta. Root administrators review the waitlist and send individual registration invitations by email.

Each registration invitation is linked to a specific email address and contains a time-limited acceptance URL such as /auth/accept-invite?token=.... Do not forward or publish an invitation link. A forwarded invitation may be revoked.

Because access is constrained by server load and operational capacity, placement on the waitlist does not guarantee immediate access or a specific activation date. Approved beta invitations may begin with a time-limited Pro grant; after that grant expires, the account returns to the free tier unless it is upgraded separately.

Appropriate use

AI4Meta may be used for legitimate research, teaching, review, and methodological development activities, including:

• Study screening and eligibility review.
• Data extraction, coding, and reconciliation.
• Meta-analysis, network meta-analysis, qualitative synthesis, and related evidence workflows.
• Documentation, provenance review, and research collaboration.

Do not use AI4Meta for unlawful activity, deceptive publication practices, fabricated evidence, automated authorship without disclosure, harassment, credential sharing, or attempts to bypass access controls. Do not use AI4Meta as the sole basis for clinical, legal, regulatory, or safety-critical decisions.

Data you provide

Depending on how you use the app, AI4Meta may store or process:

• Account details, such as email address, name, affiliation, role, ORCID, and profile links.
• Invitation request details, such as research interests, intended use, Google Scholar profile, and publication DOI information supplied for prioritization.
• Project files, including PDFs, HTML, bibliographic records, supplementary files, and manually entered study data.
• Research artifacts, including screening decisions, codebooks, extraction fields, effect sizes, qualitative codes, notes, comments, audit logs, provenance traces, and generated drafts.
• Operational records needed to secure, debug, operate, and improve the service, such as timestamps, request metadata, error traces, rate-limit events, and aggregated usage information.

Do not upload information that your institution, funder, publisher, data-use agreement, or ethics approval prohibits you from processing in third-party software. This includes confidential manuscripts, embargoed files, personal health information, identifiable participant records, sensitive personal data, or copyrighted material that you are not authorized to upload.

AI processing and third-party providers

AI4Meta includes AI-assisted features. Depending on configuration and feature choice, prompts, excerpts, metadata, tables, extracted fields, or other project content may be sent to third-party model providers or to model infrastructure configured by the user or administrator.

Provider behavior, retention, logging, and data-use terms are governed by the relevant provider account and provider policies. If your research requires a particular provider, region, retention posture, or institutional contract, configure that route before using AI-assisted features. Bring-your-own-key and local or institution-controlled model routes may reduce reliance on shared provider credentials, but they do not remove the researcher's obligation to verify compliance.

If a project contains sensitive, confidential, regulated, unpublished, or embargoed material, disable AI features or use an approved model route until your team has confirmed that the processing arrangement is acceptable.

Accuracy, provenance, and human review

AI-generated outputs are drafts. They may be incomplete, biased, outdated, internally inconsistent, statistically inappropriate, or factually wrong. They may misread tables, invent values, omit exclusions, hallucinate citations, or produce plausible but unsupported explanations.

Researchers are responsible for verifying all AI-assisted output before relying on it, sharing it, citing it, exporting it, or publishing it. At minimum:

• Confirm extracted values against the original source.
• Review eligibility decisions, labels, codes, and exclusion reasons.
• Check statistical assumptions, model choices, units, transformations, and effect-size direction.
• Preserve provenance where available, including quoted support, page locators, document identifiers, and audit trails.
• Use independent human review when required by the protocol, journal, funder, or institution.

AI4Meta is intended to make verification easier, not optional.

Privacy, confidentiality, and security

AI4Meta applies access controls, authentication, role-based permissions, logging, and operational safeguards appropriate for a beta research application. No beta system should be treated as risk-free.

You are responsible for maintaining the confidentiality of your account, invitation link, API keys, project share links, and collaborator permissions. Remove collaborators who no longer require access. Do not place secrets, private API keys, passwords, or patient identifiers into prompts, notes, titles, project descriptions, or uploaded files unless your institution has explicitly approved that use.

AI4Meta does not sell user data and does not share project data for advertising. Data may be processed to operate the app, provide support, diagnose reliability or security issues, and deliver the features requested by the user.

Research compliance and user responsibilities

Before using AI4Meta for a project, confirm that use of the software is compatible with:

• Your study protocol, preregistration, review methodology, or analysis plan.
• Institutional review board, ethics committee, or data governance requirements.
• Publisher, database, library, or copyright license terms for uploaded material.
• Funding, sponsor, or collaborator data-use obligations.
• Disclosure expectations for AI-assisted screening, extraction, synthesis, writing, or analysis.

AI4Meta does not determine authorship, resolve research misconduct questions, approve clinical recommendations, or certify that a review meets PRISMA, Cochrane, Campbell, JBI, GRADE, CONSORT, or other standards. Those responsibilities remain with the research team.

Export, correction, and deletion

Users may request access, correction, export, or deletion of account and project data. Signed-in users should use the in-app feedback or support channel so the request can be associated with the authenticated account. Other requests may be sent to support@ai4meta.net.

Some operational records may be retained for a limited period where necessary for security, abuse prevention, backup integrity, legal obligations, or auditability. Deletion of project content may not remove material that another authorized collaborator has independently exported or copied outside AI4Meta.

Changes, limitations, and no warranties

AI4Meta beta access is provided on an "as is" and "as available" basis. The service may be interrupted, delayed, rate-limited, changed, suspended, or discontinued during beta. AI4Meta does not warrant that the software is error-free, uninterrupted, suitable for a particular research purpose, accepted by a journal or regulator, or capable of detecting every error in a review.

To the maximum extent permitted by law, AI4Meta is not responsible for indirect, incidental, consequential, or special losses arising from beta use, including lost work, lost time, publication delays, incorrect analyses, unsupported conclusions, or decisions made from unverified AI-generated output.

Contact

Questions about this notice, beta access, data handling, security, or an institutional research workflow may be sent to support@ai4meta.net.

Please include enough context to identify the relevant account, project, or workflow, but do not include passwords, private API keys, patient identifiers, or confidential manuscript content in an email unless an approved secure channel has been arranged.