This week, we have decided to undertake a seemingly impossible task. We want to use machine learning to help make the lives of our video archive selection librarians easier.
Here at CNN, we have both production and archive computer servers. Content moved back and forth between these servers as it is needed. When it comes to archiving content, however, the entire process is incredibly extensive. Our archivists follow a 24-step protocol for determining what gets archived and what gets not.
My boss and I are trying to see if there is a way we can simplify this process. See if there is a way we can generate a filter on the production server content that helps block items we know we will not archive and potentially eliminate a few steps for the archival team. Now, we are very aware that our efforts may not amount to anything. When you are working for a company as large as CNN and our archived content comes from bureaus around the world, it is very difficult to ensure everyone is on the same page when it comes to documentation. Different show teams want different clips and only want their clips archived, when in reality that is just not conducive to store the same clip multiple times long-term.
The selection team is up to their noses in records every day trying to determine what to archive and what to skip. They evaluate around 3,000 records each day for archiving. We are doing everything we can to see if there is a way to simplify their methodology. Can we implement machine learning to filter out records we know for sure we will not archive? What information do we need to make that happen?
It has been a long week, for sure. We are completely changing gears from the coding I have done in the past and the materials I am working with. I am trying to compress my learning time, as there are two of us working on this project and I want to make sure I am staying on top of things.
For our intern event this week, we had a Town Hall meeting with Jeff Zucker, the current CEO of Warner Media and Sports. We had the opportunity to ask him questions about himself, the future of the company, and any other pressing questions we had. As per usual, most of the questions were journalism-based. So I felt a little out of place. We also only had an hour, so not all of us were able to ask the questions we had. If I had gotten the chance to ask mine, I would have wanted to know what his goals were for digital research and innovation internally at CNN. Everyone else is so focused on the external big picture. I have always worked behind the scenes. It would have been cool to hear what his internal goals were.
Then again, he has a background in journalism and production. My questions are a little too niche for that.