On Monday, the Los Angeles Times released a new feature that uses artificial intelligence to provide a bias rating on news stories and bullet points that offer “additional context” and opposing views.
The feature, created without the input of the paper’s editorial staff, is the “next evolution of the L.A. Times,” the paper’s owner Patrick Soon-Shiong said in a letter to its readers.
“I believe providing more varied viewpoints supports our journalistic mission and will help readers navigate the issues facing this nation,” Soon-Shiong said.
Within 24 hours, the AI feature was taken down after a New York Times reporter noticed that the tool downplayed the racism of the Ku Klux 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 note read on a story celebrating the 100-year anniversary of Anaheim residents kicking the Klan out of its local government.
The episode is emblematic of how news organizations around the globe are scrambling to hold onto consumers and remain relevant in the new digital world. Newsrooms already have endured a reduction of 30,000 jobs between 2008 and 2020, according to the Pew Research Center.
Generative AI has intensified the threat to news organizations, as it promises to answer news questions without directing any traffic to the organization that spent time, labor, and money to produce the story.
So far, at least 14 news organizations have chosen to fight AI companies like OpenAI and StabilityAI in court, claiming that their copyright was violated when millions of news articles were scraped to train generative AI programs.
But others have decided to forgo a fight and instead partner with these companies, hoping to squeeze out some form of guaranteed payment while they still can. The move has borne some initial fruit. OpenAI has agreed to help fund at least four local newsrooms as part of its deal with Axios.
At first glance, this strategy makes sense, given newspapers’ recent history of adapting to technological change. News organizations lost billions of dollars in the early days of the internet, allowing sites like Craigslist to destroy their classified sections, even as they gave away their product for free online.
But there is financial danger in partnering with tech giants. In 2016, Facebook lied to news organizations about how lucrative videos were, causing many newspapers to “pivot to video” only to lay off those workers once reality sunk in.
Facebook, now called Meta, also suckered a number of news organizations by promising to pay them for third-party fact-checking. They abandoned the move soon after President Trump was reelected, leaving those organizations once again out in the rain.
News organizations are currently stuck in a conundrum as they determine how much they should trust their likely executioner, with no easy answer.
The only hope is that going forward, whoever makes these decisions remembers three things: 1.) journalism is hard; 2.) generative AI is dumb, and 3.) please let a real human proofread summaries before you accidentally endorse the KKK.
