In early March, Ground News, a subscription-based news bias-detection platform, rolled out a new device labeled “Frames” that allows users to receive the same story from multiple perspectives quickly.
Founded in 2018 by two NASA engineers, the site had previously categorized news sources via a combination of algorithms and manual control, and sought to score articles about the same topic on their political bias, factual reliability and ownership.
“They have been doing a lot of advertising about Ground News on YouTube,” Mikael Nagy ’24 said. “I visited the site once but there [are] a lot of news bias sites. With an AI algorithm, I think it makes them [and the ratings] more trustworthy.”
Frames works by ingesting articles that AI is either left, center or right, and creating a unique bullet point summary from each perspective. According to Ground, it’s been trained to pull the most influential and well-read sources from each side and a lot of work was spent ensuring that the “key points” of each article were presented in the bullet points.
A visit to the Ground News site and an example of Frame working can be seen with a recent news report that Trump was set to announce tariffs on automobiles, with one to two sentence concise bullet points emphasizing different points of significance. Right-leaning articles discussed how Musk’s expertise at leading a multitude of companies can help the Pentagon while left-leaning articles talked about his potential conflicts of interest.
More than simply an array of bullet points, Ground News AI also presents beneath the summary a long list of sources with the AI’s classification on how the article politically leaned and whether it was primarily factual.
“For intricate topics, it seems pretty helpful,” Chirag Naik ’25 said. “Generally I visit just one or two sites to get my news, but this seems much more effective at getting a broad perspective on an issue.”
