The Los Angeles Times launched a new artificial intelligence-driven “evolution” to its company this week, with billionaire owner Dr. Patrick Soon-Shiong writing in a letter to readers that the newspaper is “beginning a new effort” to offer multiple perspectives on its opinion columns. The AI feature, known as Insights, will not only summarize data and viewpoints from a written column, but also offer online readers real-time counterpoints to that same story at the bottom of the page.

What’s more, the Insights feature will provide readers with an AI assessment of the article’s “viewpoint” — that is, whether or not the writing adheres to a right-leaning, left-leaning or centrist point of view. Insights was developed by the Los Angeles Times in collaboration with AI news summary app Particle and San Francisco-based Perplexity AI.

In just one day of activity, the Insights feature has already run into trouble. Per New York Times tech reporter Ryan Mac, an AI summary of a Feb. 25 article by columnist Gustavo Arellano offered, under its “Different views on the topic” section, a softened perspective on one of America’s longest-running and most notorious white supremacist groups, the Ku Klux Klan, and its long history in Southern California.

Arellano’s article focused on the KKK’s place in Anaheim city politics a century ago. The AI-generated alternate viewpoint downplayed the KKK’s history there, saying that “local historical accounts occasionally frame the 1920s Klan as a product of ‘white Protestant culture’ responding to societal changes rather than an explicitly hate-driven movement.” A screenshot of the summary, taken by Mac, also noted that “critics argue that focusing on past Klan influence distracts from Anaheim’s modern identity as a diverse city.”