Let’s face it, some documents are pretty boring. Official forms like tax rolls, death certificates, and prison registers don’t tell the same kind of stories that letters or diaries do. But that doesn’t mean there are no stories behind the people and events represented in these government forms. In many cases these documents might be the only record of the names of people who were too poor, marginalized or illiterate to leave their own writings.
How can we make it easier for volunteers to transcribe these sources?
We’re excited to introduce AI Drafts for field and spreadsheet based projects. If enabled by project owners, this feature lets transcribers start from an AI draft to populate the fields the project has configured, reviewing and editing from there. The feature works the same way AI Drafts do for document based collections – if an AI Draft is used, it's shown in the activity stream and the versions screen.
The only difference is that project owners need to configure the fields—as they already do—by naming and describing what to pull from the image. After that, they can kick off AI Drafts from the settings.
If you’re going to try this out, we suggest you schedule a call and work with us – we can review your field based configuration and run a page or two with you so you get feedback on how it’s working.
One of the most encouraging things we’ve observed from document based AI drafts is how much they’ve helped people transcribe harder material like account books. We’re curious how field based AI Drafts will help archives open new types of documents.
Leave a Reply
You must be logged in to post a comment.