Inside Stacker's approach to integrating AI into content distribution

The use of Generative AI has the potential to significantly enhance—or harm—newsrooms and platforms. Most organizations are grappling with how to ethically and intelligently incorporate AI into their workflows, and Stacker is no different. We’re excited to pull back the curtain and share some of the ways we’re utilizing AI.

At the heart of Stacker’s mission are quality and trust. For this reason, we’re cautious and methodical about our use of AI. Our publishing partners rely on us to provide vetted, high-quality stories from both our newsroom and outside contributors. Stories they can confidently publish without needing to further vet. That's why we have a strict policy never to use Generative AI in the creation of stories. All content is written and edited by humans.

But we believe that AI can be utilized to help with production and research tasks both inside and outside the newsroom. It can be a wonderful personal assistant that can perform specific tasks. Publishing partners and Stacker Connect contributors can receive suggestions and insights which they can choose to use or not.

Below are a few examples of how we’re currently using AI at Stacker:

Headline Brainstorming

As a content distribution platform, Stacker syndicates stories to over 3,000 publications. We’ve heard feedback from publishers that they often want to customize headlines—whether to localize them or to differentiate their story from others publishing the same wire content. In response, this week we’re releasing our first Generative AI feature (in beta) on our Story Hub platform.

When a news outlet uses Story Hub to publish a Stacker story, they now have the option to receive suggested alternate headlines that may better suit their audience. We’re partnering with Nota, one of the most exciting newsroom AI tools out there, to bring this feature to life. Importantly, in line with our AI guidelines, a human editor—in this case, from our publishing partners—makes the final decision on which headline to use. We will also be rolling the tool out internally for our own journalists to get inspiration for headline choices—though again, a Stacker editor will still have the final call on each story's framing.

Data Analysis & Content Strategy

For the past three years, every Stacker story that is syndicated to 3,000 publishers has carried a pixel that provides valuable insights into story pickups and reach. You can imagine this is a massive dataset with unlimited potential to inform our newsroom and our Stacker Connect contributors as to what our publishing network is interested in.

For example:

• Football stories perform better in August and January (just before regular season kickoff and the Super Bowl) than they do in September and February.
• Health & wellness stories are more likely to be syndicated in Fall than Summer.
• Stories categorized as News are likely to perform well any day of the week…but Entertainment and Lifestyle stories are most likely to generate high page views on Fridays.

All of these insights were generated with the support of Generative AI tools to help analyze a massive dataset and are provided to content creators to help with story scheduling. But, as always, a human is required to clean the data, establish the methodology, and—most importantly—vet the output.

Story Localization

Stacker’s newswire includes stories about all 50 states and 384 metropolitan areas. While not exactly AI, the tools we’ve created leverage automation that allow data reporters to transform large national data sets into localized state-, country-, or metro-level stories. Many other news organizations—like the Associated Press, which has used this technology since 2014—employ a similar method. Data is compiled, cleaned, and vetted by our talented team of data journalists, and all copy is written by our reporters and editors.

Where do we go from here?

We will continue to experiment with generative AI to benefit our newsroom and our partners. Newswire subscribers may see additional features added to Story Hub that help them work faster. Stacker Connect contributors can expect AI-assisted insights on which stories are likely to perform well on the wire.

Through it all, we remain committed to being transparent and cautious in our approach to AI.

Ken heads up the distribution and product teams at Stacker. He previously led product teams at The Associated Press and The Nielsen Company.

Photo Illustration by Stacker // Canva