Newspaper opinion editors push back on AI-written op-eds after WSJ defense

“If the 30-year must trade at 5.5% to clear, that isn’t a crisis. It is an invoice.”

That sentence appeared in a Wall Street Journal op-ed by billionaire financier Stanley Druckenmiller. When NOTUS reporter Jeff Stein called to ask whether a chatbot had produced it, Druckenmiller acknowledged using AI. “I write everything using AI now for the same reason I use a calculator when I do math problems,” he told Stein.

What made the episode stand out was that it was not followed by an editor’s note, a retraction or an apology. Journal editorial page editor Paul Gigot defended the piece in a follow-up column, writing, “it wouldn’t be honest for me to grandstand that our contributors never use a tool like AI.”

Nieman Lab’s Joshua Benton then contacted opinion editors at 24 prominent U.S. newspapers to see whether others would take the same approach. None of the editors who responded said they were willing to follow the Journal’s lead.

The Los Angeles Times said it allows AI for research and reporting but not generated text in outside submissions. USA Today said it accepts no AI assistance beyond basic spelling and grammar checks. Chicago Tribune editorial page editor Chris Jones put it more simply: “We are looking for human opinions, crafted and honed by humans. Period.”

Enforcing those policies is more complicated.

Denver Post editorial pages editor Megan Schrader said the paper screens letters and columns with Grammarly’s AI detector, but she is certain it has unknowingly published AI-written material. The Oregonian’s Helen Jung said she has tested submissions with AI detectors but does not find them especially reliable.

At the Pittsburgh Post-Gazette, which no longer runs an opinion section, former editorial page editor Brandon McGinley said his two-person team relied on “gut” when it suspected AI. He added that the approach might be different now if the paper still published op-eds, because AI-generated writing has become less obvious and more common.

New York Focus ran 43 recent op-eds by New York politicians through Pangram, and the majority were flagged as containing at least some AI-generated text. But a detector flag does not establish that AI necessarily wrote the material.

Jung also pointed to a policy gray area. News organizations already publish op-eds that communications staff write or heavily edit for executives and public officials. She said it would be hard to justify banning ordinary contributors from using AI to polish their own writing, though she still drew a line at having AI generate the argument or draft itself.

There is still no reliable way to prove a submission is entirely human-written. Editors can require disclosure, run detectors and look for familiar tells, but the final call remains a judgment. As Fleischer put it, readers are paying for “the words and thoughts of human beings.” Increasingly, editors have to ensure that is still what readers are getting.