Exclusive / Anti-Waymo campaign relied on op-eds written with AI - Semafor

Opponents of AI-powered cars have been using artificial intelligence to get their message out in the news media.

They’ve been waging a kitchen sink campaign, placing op-eds in local news outlets arguing the autonomous taxi company Waymo could cost jobs, endanger riders, deprive the government of information, and block fire trucks. The AI detection service Pangram found that many of the pieces were written with a significant amount of AI input, but none contain labels disclosing AI use.

Several op-eds were drafted by a PR firm, said Ira Goldstein, the former head of New York’s Black Car Fund and the author of an op-ed in the New York Daily News.

“A first draft of this OpEd was prepared by our PR firm at the time, which is standard procedure when working with a PR firm. I reviewed the OpEd, and made some edits. The PR firm then handled placement in the Daily News,” Goldstein told Semafor in a statement.

“With all that said, I stand by every word in this OpEd and believe that AI being used without my knowledge, to write a first draft of a 700-word opinion piece pales in comparison to the economic and public safety damage that thousands of AI powered robotaxis will inflict on New York City.”

(In the same statement, Goldstein said the PR firm in question later last year also dropped the campaign to work for … Waymo.)

The campaign also included series of op-eds for the Dallas Morning News on autonomous vehicle accidents by local lawyer Amy Witherite, which registered as 95% AI-written and 71% AI written, respectively, on Pangram; a San Francisco Chronicle article warning about how autonomous vehicles could go dormant in public streets during power outages, whose text Pangram reported was 88% generated by AI; another Daily News piece that Pangram said the piece was 68% AI-written; a San Diego Union-Tribune piece arguing that autonomous vehicles would likely displace jobs that registered as 95% likely AI-generated text; and a Times of San Diego piece about how the city should track Waymo’s rollout, 87% of which was likely written with AI assistance.

“Like many, many writers, I use AI tools for research, to sometimes test my arguments, help organize my thinking, and make my writing as clear as possible. These tools are a fact of modern life and modern writing,” said Michael Redmond, the San Francisco Chronicle author. Chronicle editor Matt Fleischer told Semafor that the news organization’s policies were to ask opinion writers to disclose whether they had used AI.

David Margulies, a spokesperson for Amy Witherite, the author of the Dallas Morning News article, said the firm that runs her communications helps develop, draft, edit and place these pieces, and “sometimes use AI tools to monitor media coverage, develop press lists and to prepare initial drafts, which are then reviewed, edited by our staff, fact-checked and reviewed with Amy or other attorneys in the firm before she gives her comments and final approval.”

“AI is just being used as a shiny object to try to change the subject. It won’t work. Every piece published under Amy’s name reflects her views. They are personally reviewed and approved by her before submission,” he said.

“Limited use of AI for research or polishing doesn’t automatically disqualify a submission, but we don’t accept pieces generated entirely by AI,” Dallas Morning News commentary editor Ryan Sanders said in an email.

Andrew Julien, executive editor at the Daily News, said in an email the organization aims to create stricter guardrails against AI on the site, given the publication’s ongoing litigation against Microsoft and OpenAI.

The paper feels “strongly that our readers deserve original opinions — certainly not machine-generated content created by companies we are suing over billions of dollars they owe us for stealing our content.”

The tokens burned spitting out anti-Waymo opinions are part of a monthslong battle between Waymo and its rivals over whether the autonomous driving company can operate commercially on the streets of several major cities.

The Washington Post reported this month that the Coalition for Accountability and Road Safety, a group of Waymo opponents, spent more than $2.6 million on ads in the DC area in 2026, a greater sum than the winning mayoral candidate spent in the same year. Waymo operates commercially in 15 US cities, but has faced opposition from labor unions, ride-hailing giants such as Uber and Lyft, and various AI-skeptical groups and pedestrian safety advocates.

AI writing has become a fact of life for billions of people around the world, raising questions within the media and academia about the value of and standards for human-written text.

The realm of authorship in the opinion section has also always been a bit dubious. As Semafor tech reporter JD Capeluto wrote earlier this year: “Newspaper op-eds from big names have long been ghostwritten by junior staffers or outside PR firms — no disclosure necessary, and often watered down and hyper-optimized to be as palatable as possible. In that sense, outsourcing to AI is just replacing one form of slop with another. Just look at LinkedIn, where many cringe human-generated posts have been replaced by AI.”

All that may be true. But for media organizations, allowing the proliferation of AI seems more likely to dilute, rather than enrich, the ultimate product. In a world where chatbots and agents easily answer most generic questions for free, the value of news organizations seeking payment from readers could be to provide a perspective that audiences know has been verified by a human being.

To certain journalists, there’s little point in distinguishing between AI and human authorship now that many people use AI to research and draft. In a precedent-setting move earlier this year, The Wall Street Journal’s opinion editor declared that AI-generated submissions from notable figures were fine, and didn’t need disclosure, so long as they were newsworthy, well-argued, and relevant.