Privacy notice

Who this covers

This notice describes how the FoodProt AI platform processes information about the people who use it and the research material they put into it. Accounts are created by the people who use them, with a verified e-mail address; nothing in a project is visible to anyone who has not been added to that project.

What is processed

Account information

Your email address, the identifier your account is issued under, the projects you are a member of and your role in each. This is the minimum needed to authenticate you and to enforce project boundaries.

Activity within a project

The actions you take — searches run, screening labels, extraction runs started, rows approved, corrected or rejected, configuration changes — recorded with your identity and a timestamp. This record is the audit trail, and it is the point of the product; it cannot be disabled without making the resulting dataset undefendable.

Research material you upload or fetch

Bibliographic records, article PDFs, supplementary files, parsed text, extracted measurements and the evidence snippets attached to them. This material is scientific literature and data about it, and is not expected to contain personal information beyond the authorship metadata published with the articles themselves.

Technical information

Session cookies used to keep you signed in, and server logs recording requests for security and reliability purposes. FoodProt AI does not use advertising cookies, and the public pages of this site set no cookies at all.

Other services involved

Running FoodProt AI involves a small number of other services. Which ones apply depends on how the platform has been set up by the organisation that runs it:

• Storage and sign-in. Accounts, projects, files and all project data are held by a hosted database and sign-in service on the platform’s behalf.
• Literature databases. A search sends your query to the databases you selected — for example PubMed, OpenAlex, Crossref, Semantic Scholar and Europe PMC — and open-access status is looked up through Unpaywall. These requests contain your search terms, not your identity.
• Suggestion services. If the platform has automated suggestions switched on, the text of a paper may be sent to a language-model service to suggest a screening decision or a value. Every such result is marked as a suggestion and is decided by a person. If suggestions are not switched on, no paper text is sent.

Legal basis and purpose

Account and activity data are processed to operate the platform, to enforce access control and to maintain the audit trail that gives a dataset its provenance. Research material is processed to carry out the study the project was created for, on the instructions of the organisation that holds the project.

Retention

Project data is retained for as long as the project exists, because deleting part of an audit trail would silently invalidate the dataset built on it. When a project is deleted, its records, files and run history are deleted with it. Account records are retained while the account is active and removed on request once it is not, subject to any audit entries that must remain attached to datasets still in use — in which case the entry retains the identity that made the decision, which is the whole purpose of recording it.

Access, correction and export

You can see the data held in any project you are a member of through the application itself, and export a project’s verified dataset with its audit columns. For account-level requests — access, correction, deletion, or a copy of what is held about you — contact the address below.

Assistant conversations

Questions you ask the assistant, and its answers, are stored so that you can return to a conversation. They are visible only to you, are kept separately for each project, can be cleared at any time, and are removed automatically after 90 days.

Security

Access is checked for every request against your project memberships, so an account that has not been added to a project cannot read it. Files are stored privately and are downloaded through links that expire. Credentials for connected services are held on the server and are never sent to your browser. Data is encrypted in transit. FoodProt AI does not claim any formal security certification.

Changes

Material changes to this notice will be reflected in the review date at the top of the page. This notice is not a contract; the terms of use describe the agreement.

Contact

Privacy questions, and requests concerning your own data, go to the organisation that runs this platform (rroni0528@sdsu.edu).