Vetra

Legal

AI & Data Processing

This page explains how Vetra uses artificial intelligence to analyse scientific articles, which information is involved and what the service retains.

Last updated: 2026

Essential information

Vetra uses artificial intelligence to help analyse scientific articles. Processing is limited to providing the requested features, while the project retains the results and evidence needed to continue the work.

1. Information Vetra processes

Vetra is designed to work with published scientific articles and the information needed to conduct a systematic review. Depending on the features used, processing may include:

  • Scientific articles, including the documents the user chooses to analyse.
  • Bibliographic metadata, such as titles, authors, DOIs and other information linked to the project.
  • Review instructions and criteria, including questions, selection criteria and guidance provided by the research team.
  • Work results, such as extractions, scales, assessments, decisions, conversations and references to evidence.
  • Account and usage data, needed to manage permissions, consumption, operation and security, and handled separately from scientific content whenever possible.

Permitted content. You must not upload medical records, participant datasets, confidential patient information or other personal data that can identify participants. Vetra is intended for the analysis of published scientific articles.

Responsibility for licences. By providing an article, the user represents that they hold the rights, licences, permissions or authorisations needed to access and analyse it. The user is responsible for complying with any applicable publisher, institutional or other licence terms.

2. How an article is analysed

When you use an analysis feature on an article, the process follows this sequence:

  1. 1ProvisionThe user provides a published scientific article.
  2. 2AI-assisted analysisVetra analyses the content through the artificial intelligence services needed for the task.
  3. 3ResultsStructured results and references to the supporting evidence are generated.
  4. 4CompletionTemporary processing of the article ends once the analysis is complete.
  5. 5Project continuityThe project retains the results, evidence and AI or manual origin status.

Vetra does not store or distribute the article provided: it uses it only to perform the requested analysis. If you need to analyse it again, you may have to provide it again.

3. AI providers and data protection

Vetra uses OpenAI as a provider for certain artificial intelligence features. Content is processed only to the extent needed to provide the requested analysis.

Vetra applies Zero Data Retention (ZDR) to compatible workflows and reviews the conditions of the services it uses before adding new features.

No training or secondary uses. Vetra does not use uploaded articles to train models, develop products based on their content or carry out commercial activities unrelated to the requested service.

4. Data separation

Vetra limits the information sent to each feature and avoids including identity data in model instructions where it is not needed.

  • Scientific content: the article and content required to perform the requested analysis.
  • Scientific metadata: the title, authors, DOI and other bibliographic information linked to the project.
  • Project data: the criteria, instructions, decisions, results and evidence used to continue the review.
  • User data: the email address, identifier, organisation, role, activity and consumption needed to provide and protect the service.

Personal data associated with the account is described in more detail in the Privacy Policy.

5. Processing, storage and deletion

Transient processing

The content needed to generate the analysis is processed temporarily. Vetra does not store or distribute the article provided, or add its content to the project's permanent storage.

Persistent project information

While the project remains active, Vetra retains structured results, scales, extractions, decisions, conversations, evidence, bibliographic metadata and the links needed to maintain the state of the work.

Deletion

Deleting a project starts the operational deletion of this information, which is completed within 30 days. Associated backups expire within 90 days.

These periods may be extended where required by law, necessary to prevent fraud or essential for the establishment, exercise or defence of legal claims.

6. Human oversight, errors and traceability

Results generated or assisted by artificial intelligence may contain errors, omissions, incorrect interpretations or hallucinations. The user must review the evidence and validate any material methodological, scientific, editorial or clinical decision.

Vetra is a research support tool. It does not replace professional judgement, independent methodological review, or clinical, legal or regulatory advice.

Each result should indicate whether it comes from AI or was updated manually. Where linked evidence is available, Vetra makes it easier to review that evidence in the article. This does not mean that a complete history of every change, author and timestamp is available.

7. Providers, changes and contact

Certain features depend on external infrastructure and artificial intelligence providers. Their incidents, changes or unavailability may temporarily affect the service. Vetra selects and configures those providers according to its security and privacy requirements, but does not guarantee uninterrupted availability.

This information may be updated to reflect legal, technical or service changes. Where a change is material, Vetra will provide notice through appropriate channels before it applies whenever reasonably possible.

For any enquiry about artificial intelligence, privacy or data processing, write to info@vetralab.org.

This page is also available in Spanish. If there is any discrepancy between the versions, the Spanish version prevails.