It didn't take long for the L.A. Times' new AI to defend the KKK

As hard as it is to believe, the L.A. Times’ controversial new AI pivoted to racism within 24 hours of launch. 

By Matt Schimkowitz  |  March 4, 2025 | 8:09pm
Photo by Kirby Lee (Getty Images)
News News A.I.
It didn't take long for the L.A. Times' new AI to defend the KKK
Introducing Endless Mode: A New Games & Anime Site from Paste

Who could’ve seen this coming? Less than 24 hours after deployment, the Los Angeles Times’ new artificial intelligence is defending the Ku Klux Klan, per The Daily Beast. Designed to give a full range of opinions and perspectives, Insights, developed by the company Particle, was announced yesterday to ensure the paper’s supposed biases were counterbalanced with a list of counterarguments. For example, if the article is about why the Nazis are evil, the AI will retort, “Say what you like about the tenets of National Socialism, at least it’s an ethos.” That’s pretty much what happened today in a column by Gustavo Arellano about the 100th anniversary of Anaheim removing four KKK members from its city council. Desperate to provide a counterargument, the AI opted to defend the Klan.

“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, minimizing its ideological threat,” the AI offered. The “insight” has since been removed from the article, but the bot continues still terrorizing other pieces, awaiting another chance to provide a racist perspective.

New York Times reporter Ryan Mac caught this latest instance of slop entering the news media. “Earlier today the LA Times had AI-generated counterpoints to a column from [Gustavor Arellano]. His piece argued that Anaheim, the city he grew up in, should not forget its KKK past,” Mac wrote on Bluesky. “The AI’ well, actually’-ed the KKK. It has since been taken off the piece.”

We should be shocked, but this is precisely what the L.A. Times Guild predicted would happen a day ago when the feature launched. “This tool risks further eroding confidence in the news,” L.A. Times Guild vice chair Matt Hamilton said yesterday. “And the money for this endeavor could have been directed elsewhere: supporting our journalists on the ground who have had no cost-of-living increase since 2021.”

Our sympathies are with the writers and editors navigating this minefield unnecessarily created by the paper’s owner, Dr. Patrick Soon-Shiong, who reportedly cost the paper 7,000 subscribers after he blocked the paper’s endorsement of Kamala Harris last fall. However, for a biotech billionaire and owner of one of the country’s most important newspapers, one would assume he’d ensure his junk widget worked before implementation. But this is AI we’re talking about, and the chance to replace labor with a crappy, hallucinating, racist bot is too alluring for the billionaire class. While it might’ve worked for Facebook 15 years ago, “move fast and break things” isn’t a viable mission statement for a newspaper. Maybe “move fast and fact check” would work better.

 
Join the discussion...