Sign up for the daily CJR newsletter.

On Monday, when the Wall Street Journal published an opinion piece from Stanley Druckenmiller, Paul Gigot, the editorial page editor, hoped it would make news. Druckenmiller, a billionaire investor, had been a mentor to Scott Bessent, the treasury secretary, early in his hedge fund career. Now, in the piece, Druckenmiller was essentially rebuking Bessent’s bond buyback plan. But the message was soon overshadowed by the means of its production. 

Readers on social media quickly noticed that the essay bore all the hallmarks of AI writing and posted results from Pangram, an AI detection tool, that appeared to confirm their suspicions. By Tuesday, Druckenmiller confirmed the speculation to a reporter at NOTUS. “Of course I used AI,” he said. “I’m not embarrassed by it,” he continued. “I write everything using AI now for the same reason I use a calculator when I do math problems.”

I don’t fault Druckenmiller for his use of AI. He likely assumed that if that were a problem, an editor would have flagged it—and it was pretty obvious in the writing. (“Volatility was contained, and trading was orderly—not a malfunction but the machine doing its job.”) Nor is this the first time AI has been used on an op-ed page. But the calculator analogy was flawed. A calculator can perform a discrete task, but it cannot build a complex financial model or decide what assumptions to make. The same could be said of how AI writes. 

In a statement to Semafor’s Max Tani, Gigot defended the piece. “AI is a fact of modern life. People will use it to assist in their work and their writing, including with research, checking grammar, editing and more,” Gigot wrote. “In Stan Druckenmiller’s case, we have had a relationship with him for many years, and nobody can doubt that his op-ed is his genuine opinion.”

It was a surprising response given the Journal’s stated policy on AI, which states that it may be used in specific cases such as complex data-driven investigations, article summaries, and translations, and that “in all instances, we adhere to our long-held values of accuracy, accountability, originality, quality, transparency and respect for intellectual property.” I’m not sure how one could argue that this piece met the standards of originality, quality, or transparency, since there was no disclosure of AI use.

Gigot did not respond to my request for further comment, but he did write a piece of his own, deriding journalists who had covered the topic as an “internet ethicist,” a “media hall monitor,” “scolding scribes,” and “pharisees.” Between barbs, he laid out an AI policy for the editorial page: “The rule for those of us at WSJ Opinion is that everything my colleagues and I write must be original work. No letting AI draft a column or editorial.” But “the question is more complicated for outside contributors” who “aren’t professional writers” and may have otherwise used ghostwriters anyway. 

Is the ethical concern more complicated for outside contributors? Powerful people who are given space on opinion pages do often employ help from professional writers, but a good ghostwriter isn’t just a scribe. They challenge and sharpen ideas, improving not just the writing but the essential argument. “AI does exactly the opposite,” Michael Socolow, a media scholar at the University of Maine who has written about the history of op-ed pages, said. “It doesn’t make it idiosyncratic. It flattens it and makes it interchangeable.”

As Socolow told me, the opinion section of the Journal is “being enormously devalued by making it fungible, by making an industrial product that can be made completely divorced of human intervention.” Readers of the Journal presumably turn to it because they expect a quality of thinking they can’t get by plugging prompts into an LLM. If all that matters is that a piece represents a “genuine opinion” and that “the author has the standing and credibility to make it,” as Gigot puts it, then why not dispense with the writing entirely? Interviewing Druckenmiller would have gotten the idea across and spared us the slop.

In his essay, Gigot argues that all of this is so much sound and fury, and “shows how little the press corps has thought about the reality of AI, and how little they understand the role of the opinion pages in a publication.” Perhaps that’s true! Perhaps Gigot is the only person among us who has done any real thinking on this topic. Let me turn to something that may have more authority: ChatGPT. What does it think about the use of AI in writing op-ed pieces? “For op-eds, I’m generally more skeptical than I am about AI being used for things like transcription, research assistance, or copyediting. An op-ed is supposed to be a person’s argument. Not just the conclusion, but the thinking that gets you there: the choices about what matters, the connections you make, the examples you use, the voice and rhetorical instincts that make the reader understand how you see the world. If AI is generating substantial portions of that, I think there is a real authorship problem even if the human agrees with every word.” Maybe Gigot would have preferred we just publish that.

Last April, Katie Worth, writing for the Hechinger Report, and Jessica Meszaros, of WUSF, West Central Florida’s NPR station, submitted public records requests for correspondence between the Florida Department of Education and textbook publishers. Back in 2024, when Ron DeSantis signed a bill stripping the phrase “climate change” from state law, textbook authors told the Orlando Sentinel that the term had to be removed from science books before they could be accepted in public schools. Now Worth and Meszaros wanted to follow up, to determine the extent of the revisions. For months, the department ignored their requests. “We had a plan that at least once a week or every two weeks, we would check in with the public records department, whether that was email or calling,” Meszaros said. In July, their persistence finally paid off.

Their reporting uncovered that Florida education officials “had made extraordinary demands to change climate content.” In one instance, a reference to climate change “killing at least 150,000 people per year” was changed to “human health will also be affected by the increase in CO2.” Other requested revisions included replacing “climate change” with “changes in the environment,” removing references to human activity and pollution, and altering or eliminating passages about the effects of the climate crisis. Worth and Meszaros also discovered that books most likely to mention climate were approved by state officials at lower rates and were often accepted only after revisions.

“We’re in a state that is arguably the most affected by climate change, and will continue to be. And DeSantis is just pretending like it doesn’t exist,” a middle school science teacher from Central Florida told Worth and Meszaros. Other teachers told them that, even though mentioning climate change is not explicitly banned, they are operating in a “culture of caution,” where they are told by “superiors to refrain from discussing any topics deemed ‘controversial’ for their own benefit.”

Both Worth and Meszaros told me that removing scientifically sound information from teaching materials ultimately leaves students less prepared for the future. “We really owe young people information about this world that we are handing them,” Worth said. “The effect of censoring the textbooks and keeping kids away from basic information about the climate crisis is that they’re not prepared to be citizens in the world. They’re not prepared for what is coming. And that feels deeply unfair.”

When officials broke ground on the Joliet Treatment Center, a facility for mentally ill inmates, in 2017, the director of the Illinois Department of Corrections said: “You will look at this in the next fifty years as the place that helped turn the Illinois Department of Corrections around from being a punitive place to being a restorative-justice operation.” Less than ten years later, the reality bears no resemblance to that vision.

In a damning investigation for the Marshall Project, WBEZ, and the Chicago Sun-Times, Shannon Heffernan found that a prison designed to help people with serious mental illness “has become a place of such extreme desperation that people set their cells—and themselves—on fire.” Joliet houses about two hundred people. According to Heffernan’s report, there have been more than seventy fires at the center in the past three years, including a dozen cases in which men set themselves ablaze and two deaths. Last year, there were roughly three fires a month. “When I saw the number, I was like, ‘Oh wow, that is a lot,’” Heffernan told me. 

In 2024, when a man set fire to his cell, guards sprayed a fire extinguisher and pepper spray through a slot in his door and left him locked in for half an hour. When firefighters arrived to clear the smoke, guards did not tell them that an inmate was inside—breathing but unconscious. Even after the firefighters found him, the guards refused to let an ambulance take him to a hospital for almost twenty-eight more minutes, insisting they had to finish filling out paperwork. Two weeks later, the man died from his injuries. 

“Officials at the Joliet Treatment Center have repeatedly failed to meet their responsibilities to prevent or respond to the fires,” Heffernan writes, citing government records. “Cellblocks lacked sprinklers, and fire alarms were turned off or inoperable. At an annual inspection, the prison would not provide a fire evacuation plan.” Since the investigation was published, several lawmakers have called for reforms, though no changes have been implemented. 

Heffernan, who has covered prisons and jails extensively, told me that when she first got a tip about a fire at Joliet, early last year, she wasn’t immediately convinced there was a story, but she made some initial public records requests anyway. “When you report on something where there’s as many horrors as there are in prison, you have to be very thoughtful about which horrors you’re going to follow,” she said. “One big lesson for me, which is a lesson I have to learn over and over and over again, is the thing as a reporter that may not shock you anymore shocks your audience.”

If you have a suggestion for this column, please send it to laurelsanddarts@cjr.org. We can’t acknowledge all submissions, but we will mention you if we use your idea. For more on Laurels and Darts, please click here. To receive this and other CJR newsletters in your inbox, please click here.

Has America ever needed a media defender more than now? Help us by joining CJR today.