New York politicians are using AI to write their op-eds

This article was originally published by New York Focus.

If you’ve read an op-ed attributed to a New York politician in the last few months, there’s a good chance that not all of it was written by a human being.

To track the rising use of artificial intelligence in political communications, New York Focus analyzed 54 opinion articles attributed to New York elected officials that have been published since May 1 by three prominent news outlets: City & State, the Albany Times Union, and the New York Daily News. We ran each of those op-eds through the AI-detection tool Pangram and contacted the officials whose pieces were flagged as containing AI content.

Nine politicians — including the governor’s running mate, the mayor of Albany, and the chair of the Senate technology committee — acknowledged that they or a co-author used AI in producing 10 different articles. Another five said they could not be sure whether AI was used, four denied using AI, and nine did not respond to requests for comment. Twenty-six of the articles were classified by Pangram as fully human-written.

Just over half of the pieces — 28 in total — included text that Pangram identified as AI-generated.

None of the pieces disclosed the use of AI — even though lawmakers passed a bill this year that, if signed by the governor, would require news outlets to disclose substantial AI content. (One outlet, City & State, added disclosures to some articles after receiving New York Focus’s questions.)

The findings show how AI is already shaping a critical part of politicians’ jobs: the way they communicate with the public. These instances of AI writing are harder to track than some previous cases, like the uncanny fake videos released by Andrew Cuomo in last year’s New York City mayoral race and by Bruce Blakeman in this year’s contest for governor.

Elected officials have long relied on staff to ghostwrite their op-eds, a practice that many of the officials reached by New York Focus were quick to note. Relying on AI tools could be viewed as just another form of outsourcing. And about half of U.S. adults now say they use AI chatbots, according to a June survey by Pew Research Center.

But the public has reacted with outrage to even minor uses of AI in politics. Cuomo’s mayoral campaign faced an uproar last year over a single citation in his housing plan that revealed the use of ChatGPT for research, even as his aides insisted that AI was not used as a writing tool.

Adrienne Adams, the Democratic nominee for lieutenant governor and former New York City Council speaker, acknowledged through a spokesperson that AI was used to compose her Daily News op-ed from May touting her record on child care. The spokesperson said that after Adams joined Hochul’s re-election ticket, Adams’s team used AI to compile and summarize her public comments and transcripts from internal interviews about child care and family issues.

Albany Mayor Dorcey Applyrs acknowledged that her office used AI to help write her July op-ed in the Times Union about the city’s fiscal crisis. Her spokesperson Zipporah Hommel did not give details about how AI was used, but said the mayor’s office used “a host of digital tools, including generative AI, to support synthesizing, editing, and proofreading content.”

“All content is vetted and fact-checked by members of the team for accuracy and cohesiveness before being distributed,” Hommel said.

Frank Morano, a Republican City Council member from Staten Island, confirmed that he used ChatGPT to compose a City & State op-ed published in May about legalizing bodega cats, co-bylined with State Assemblymember Keith Powers. Morano said he used AI to “organize and refine” his ideas, but emphasized that he and his staff edited the piece.

“I don’t view AI as fundamentally different from other technologies that have changed how people work,” Morano said in an interview. “The question is, are you representing someone else’s views as your own? These are 100 percent my positions and AI helps me organize my ideas and communicate more efficiently sometimes.”

But the stakes of publishing AI-generated content could soon get higher. In June, the state legislature overwhelmingly passed a bill that would require news outlets to add disclaimers to any published content “substantially composed, authored or otherwise created” through the use of generative AI. If an outlet publishes AI content without that disclosure, the attorney general would be empowered to investigate and seek penalties starting at $1,000. 

The so-called FAIR News Act would likely apply to op-eds submitted by outside contributors like elected officials, according to both its supporters and critics.

“The public needs to understand that something was largely generated,” said state Senator Patricia Fahy, the bill’s prime sponsor. “I just think we’re at the beginning of trying to regulate this.”

Hochul has not taken a position on the bill, and her office said only that she would review it. The FAIR News Act was supported by labor unions but opposed by news industry groups like the New York News Publishers Association, which argues that the bill would violate the First Amendment, risk politicized investigations against news outlets, and create too much uncertainty about what constitutes “substantially” AI-generated content. (New York Focus is a member of the News Publishers Association.)

Most of the officials contacted by New York Focus said their AI use was limited; none admitted to fully AI-generating any piece of writing, although some were identified by Pangram as extensively AI-written.

City Councilmember Harvey Epstein’s office suggested its staff would reevaluate their use of AI after acknowledging to New York Focus that the technology was used in a pair of City & State op-eds on horse-drawn carriages and after-school programming that he co-authored with colleagues in the City Council and state legislature. Epstein spokesperson Jackson Fischer-Ward said both pieces were “conceived and written by our communications team.”

“As part of the editing and proofreading process, AI tools were used in a supplemental way. We’re evaluating if AI tools will be part of the editing and proofreading process in future written communications,” Fischer-Ward added.

Some officials said AI made its way into their op-eds through co-authors and group writing efforts. Councilmember Julie Won acknowledged using AI in a Daily News op-ed from June about her campaign for Congress, telling New York Focus that a three-person team used AI “to help refine and suggest edits for our final draft,” but adding that “the ideas and substance are our own.” 

“And as the op-ed reads, my voice and my opinions are captured by my team from multiple sessions of brainstorming, reviews, and edits, and as our final step we had AI suggest edits,” Won said. “As an immigrant and second language speaker, the tools help refine the final product but it does not have the capability to replace myself and my team for my identity and our campaign’s mission.”

AI was also used in a City & State op-ed about the city’s waterways co-authored by Councilmember Virginia Maloney and Julie Raskin, leader of the nonprofit Waterfront Alliance, according to Raskin.

“The op-ed was written by staff and AI was used to edit for clarity, in accordance with the Waterfront Alliance’s generative AI policy,” Raskin said in a statement. She added that the piece was then edited by staff of the Waterfront Alliance and Maloney’s office.

And state Senator Kristen Gonzalez, a leading advocate for AI regulation and chair of the chamber’s technology committee, confirmed that AI was used in an op-ed she bylined in City & State calling for data-center regulation. Gonzalez’s spokesperson, Kat Deus Veloso, said Gonzalez’s office did not use AI, but that one of several “stakeholders” that contributed to the article acknowledged “using AI to assist them.” (Veloso declined to name that person. The Pangram analysis identified only one paragraph as AI-generated, a much smaller share than in most of the other pieces.)

Other officials called their use of AI a legitimate writing tool, in moderation. Councilmember Shahana Hanif said her office used AI to help write her City & State op-ed about mass surveillance at Madison Square Garden, co-bylined with a surveillance oversight advocate. 

“My team and I have used AI tools to perform office functions, including limited writing and editing, and to make our work more efficient,” Hanif said in a statement, adding that the technology should be used responsibly and regulated to prevent abuse.

And a staffer for Assemblymember Landon Dais said his office used AI to edit his Daily News op-ed about deed theft after the newspaper asked them to shorten the piece by more than 800 words to fit in the print edition. Dais’s chief of staff, Quintin Maidment, said it was not the first time his office had used the technology.

“In general, we use AI to make content clearer, to analyze research, to provide translation and to design flyers,” Maidment said in an email. “This allows our staff to do their jobs more effectively and helps a small office handle a very large volume of work.”

Five other officials said they did not know whether AI was used because so many people had a hand in writing a given op-ed.

Reached for comment, editors for the three publications expressed varying levels of concern.

Peter Sterne, news and opinion editor for City & State, said the use of AI violated the publication’s policies. Following New York Focus’s inquiry, the publication added disclaimers to the five articles whose authors acknowledged AI use.

“In the past, we have relied on the honor system, but going forward, we plan to run every op-ed submission through AI detection tools,” Sterne said in an email. (Sterne previously worked at New York Focus.)

Andrew Julien, executive editor of the Daily News, said in an email: “We appreciate the information provided by New York Focus. We will review the op-eds in question along with our processes regarding disclosure of AI use to assure compliance with internal policies and best practices.”

Akum Norder, senior editor for opinion at the Times Union, said in an email that the newspaper expects all submitted content to be “original.” The publication is “not OK” with contributors using AI to do their writing for them, but finds it acceptable to use AI for editing and proofreading.

Norder said she had not suspected AI authorship in Albany Mayor Applyrs’s op-ed because an initial draft contained punctuation errors and because the piece resembled the writing style of four previous pieces that Applyrs had submitted to the newspaper.

“AI is only going to get better at writing and harder to detect,” Norder said. “As technology evolves, our procedures will evolve with them, as best they can.”

Camille Rivera, a progressive Democratic strategist, said she recently rejected a client’s suggestion to use AI to compose a written piece as part of a political campaign, worrying it would compromise the candidate’s integrity.

“We’re not going to run an AI campaign. That’s not what voters want,” she recalled saying. “They want to see what your real thoughts are.”