AI feature makes SDR article deposits even easier

Back in April, we announced the release of a simplified, faster workflow for depositing articles into the SDR (Stanford Digital Repository). We've now enhanced this already streamlined workflow with an AI-enabled feature that can add the article's abstract to the metadata. With one click, the new feature helps to make the metadata for your article deposit more complete, improving the chances that it will be identified in web searches.

To use the articles workflow to deposit an open access copy of a previously published work, an SDR depositor selects the "Deposit article by DOI, PMCID" button on their SDR dashboard, uploads the article file, and enters the DOI (Digital Object Identifier) or PMID (PubMed ID) for the already-published version of the work. We retrieve existing metadata about the work so that no manual entry of metadata is required.

The new AI feature comes into play when the existing metadata does not include an abstract. In these cases, depositors will see a new button that says, "Get abstract from PDF using AI." Clicking this button will retrieve the abstract from the file you uploaded and include it in the metadata for your deposit. Easy as that!

The new feature works in a closed Stanford environment and none of the submitted data is used to train models. The abstract is being pulled from the submitted file; the AI tool does NOT attempt to write its own abstract. Once the abstract is retrieved, you'll be able to edit it to correct any mistakes, or delete it entirely if you decide you don't want to use this feature.

As always, we're eager to hear your feedback about this latest improvement, which you can provide using the "Contact SDR staff" link on any page at sdr.stanford.edu.

If you're a Stanford author interested in depositing open access copies of your publications in the SDR, please contact us using our web form.