Syracuse officials are using AI to varying degrees. Central Current set out to figure out the city's plan for regulating AI use by its employees and the Common Council members' AI use. Credit: Michelle Gabel | Central Current

On April 30, Mayor Sharon Owens and Syracuse lawmakers received from Don Johnson, chairman of the Citizen Review Board, an email containing two documents for their review.

One of the documents, a legal brief, declared it was submitted by the city of Syracuse’s lawyers. The brief’s contents, though, challenge city lawyers’ official opinion on a recent ruling of the New York’s Fourth Appellate Division to pause Rochester’s independent police misconduct review board. 

Elected officials, activists, and lawyers in Syracuse are at odds over how to proceed with civilian-led police accountability efforts. The mayor’s lawyers believe the Rochester ruling applies to Syracuse’s review body, too. 

Johnson, also a leader within the local chapter of the NAACP, in the email stated he had sought outside legal counsel and found legal advice contradicting the conclusions of the city’s counsel. He asserted the documents refuted the city’s lawyers, and implored the mayor and councilors to either read the documents or listen to an “audio analysis” Johnson had also compiled and shared in the email.

“Taken together, these documents make the case that the Syracuse CRB stands on legally distinct ground,” Johnson wrote. “The precedent does not control us.”

The city’s lawyers did not commission or produce the documents, however — Johnson and artificial intelligence did. 

The cache of documents caught lawmakers off guard. At the bottom of one page of the brief, which includes a mock up of an oral argument, italicized text reads “In the Voice and Method of Thurgood Marshall”— the U.S. Supreme Court’s first African-American justice. Before joining that body, Marshall in 1954 on behalf of the NAACP successfully argued that “separate but equal” policies were unconstitutional in the landmark Supreme Court case Brown v. Board of Education.

“It’s saying okay, I’m looking at how Thurgood Marshall would have written this documentation,” Johnson said. “I’m going to write it in that same style. That’s all it is doing.”

A Central Current reporter asked Johnson if he believed the AI-generated statement was an affront to Marshall’s memory, given it implied that a machine could match the renowned attorney’s rhetoric.

“Machines can do more than match it,” Johnson said. “Machines can do what Thurgood would never be able to do: look at thousands of documents and come up with a conclusion.”

Central Current asked why Johnson didn’t notify the recipients of his email that he had used AI to generate the materials he shared.

“I don’t think I have to,” Johnson said.

As Syracusans grow more attuned to the private sector’s increasing use of AI, Central Current asked the city and Common Council if and how elected officials and their staffers are using AI.

In a statement to Central Current, Mayor Sharon Owens said AI represents an opportunity to enhance municipal operations, so long as it is implemented in full accordance with city policies. Owens stressed the importance of maintaining public trust while trying to leverage AI.

“Our responsibility extends beyond innovation; it includes ensuring that any use of AI remains transparent, ethical, and aligned with the best interests of our constituents,” Owens said.

In Central Current’s 2025 Democratic primary debate, Owens was the only of three candidates to rule out using generative AI in official city correspondence. The city has maintained Owens’ pledge so far, said city spokesperson Sol Muñoz.

Regulating and leveraging AI in government operations is a challenge across the country. Sarah Hubbard, associate director for Technology and Democracy at Harvard Kennedy School, has studied the ethical integration of AI in local government. Some of Hubbard’s research delves into mitigating potential threats AI poses to democracy and the public’s trust of its representatives.

Researchers have so far struggled to establish a one-size-fits-all standard for the technology. 

Differing municipal priorities and the speed with which the AI industry is evolving play a part in the struggle. 

“Local leaders have to be really careful to try to maintain and build that trust when using technology,” Hubbard said. “And that is … uncharted territory.” 

‘Human in the loop’

As the city considers how and where the implementation of AI may improve its practices, “augmentation” is a guiding word, said the city’s director of digital services, Vinny Scipione.

Rather than looking for areas where AI can automate government services or replace human employees, Scipione said, the city is exploring how city employees might apply AI to make daily workloads more effective and efficient.

Every city employee who uses a computer in their everyday work has access to Microsoft Copilot or Copilot Chat, Scipione said, which he believes can offer basic but valuable analytical assistance.

“If you’ve got, you know, a simple set of data or other tasks that you’re looking to get assistance with, it’s a tool available that you can use for those purposes,” Scipione said.

The city, Scipione said, could best be described as “intrigued” in its current perception of AI and its applicability in local governance. The city is not currently instructing employees to use AI in their work if they do not want to, Scipione said, but is encouraging staffers to experiment with the AI tools at their disposal if they so choose.

The city’s data governance committee is also drafting department-wide guidance aimed at teaching employees “literacy” with AI, he said. The goal of that education is to help employees learn the basics of engaging with AI and encourage those curious on how to craft prompts to direct the AI tool.

As the city plots its guidance and oversight measures for AI, it remains unclear if Owens is considering instituting a holistic policy governing AI usage across all departments. Central Current asked Owens if she is considering such a policy, and if so, what it would entail, but Owens did not directly address the question in her response.

Scipione, though, said that the city already has oversight of its AI usage through multiple overlapping policies regarding data collection and use. The data governance policies are a few years old because the city tried to create data policies before AI usage became more prevalent, Scipione. 

The city is exploring using AI to speed up its permitting process and in July applied for a grant through the U.S. Department of Housing and Urban Development that would pay for the effort.

“There should always be a human in the loop, and any process should not be used for authoritative decision making,” he said.

Despite the city seeing broad potential in AI, Owens’ communications office wants to maintain the “human element” in communications and is “very skeptical” about using AI, said Muñoz.

The communications team still produces the content for their constituent audience without the use of AI, Muñoz said, because they believe that a machine cannot communicate information as accurately and authentically as humans can.

“Who knows the residents in the city better than us?” she said.

‘You can’t let a machine think for you’

AI has entered the Common Council’s chambers, too, but the council is not currently exploring structured implementation akin to the mayor’s office’s endeavors.

Common Council President Rita Paniagua in an interview with Central Current said she does not frequently use AI but knows that some councilors do. The council has not held any big-picture discussions on AI, but Paniagua said she believes that an overarching policy would be productive to ensure the city wasn’t abdicating responsibilities to a bot.

“You can’t let a machine think for you, and that’s my stance on it,” Paniagua said.

Paniagua said she only uses AI to check her grammar before making public statements or issuing correspondence. Paniagua, who is bilingual, said AI can also aid in translation but warned that the tool is not sophisticated enough to translate with precision.

“Even when you’re translating, you need to go back. You need to read, and you need to correct, because it’s a machine,” Paniagua said. “It’s not interpreting. It’s just doing it literally, and it comes out with as many errors as you can think of.”

That limited approach appears to be shared by the councilors Paniagua leads. Four councilors described in interviews with Central Current using AI mainly for grammar checks and broad research of new topics. Two others said they don’t use it at all.

Councilors’ uses of AI vary: Councilor Corey Williams has used it sparingly for preliminary research on legislative proposals. Councilor Hanah Ehrenreich sees ChatGPT as a “second editor tool” that can add clarity or vibrance to her writing. Ehrereich does not use AI to develop legislation, she said. 

Councilor Donna Moore explained that she used ChatGPT to help her finalize a letter to the editor of syracuse.com | Post Standard advocating for the city to allow the Citizen Review Board to continue its operations.

“The substance of the letter was mine. I just asked it to fix my grammar and my spelling,” Moore said.

Moore said generative AI should “probably” not be integrated into city communications.

Councilors Helen Hudson and Jimmy Monto, the president pro tempore of the council, were the least enthused with AI among their peers. Hudson doesn’t use AI herself, but said she understands that it may have the potential to improve governmental processes in the future. She recommended her colleagues and the city approach AI with healthy skepticism as they consider how AI implementation could help continue modernizing the city government.

“I got elected as myself, I don’t need a computer to speak for me,” Monto said.

Like Hudson, Monto acknowledged that AI may offer potential benefits to the Syracuse government, but said he is very concerned about AI reducing the city’s workforce.

The councilors agreed: AI has utility, but only if used with scrutiny.  

Councilors Rasheada Caldwell, Chol Majok and Jones-Rowser did not respond to several phone calls and text messages from Central Current reporters requesting interviews or comment for this story.

Jones-Rowser, who has often promoted AI content on her Facebook page, has used AI in the past to justify her position on contentious legislation. In February, before the Common Council deadlocked a much-anticipated vote on Good Cause Eviction, Jones-Rowser prefaced the vote with a statement that made several references to “the data.”

A Central Current reporter contacted Jones-Rowser later that day to ask for the data that undergirded Jones-Rowser’s staunch stance against enacting the tenant protections legislation. Among other pieces of information, Jones-Rowser sent three screenshots of AI-generated responses from Google’s Gemini to simple queries like “why do landlords leave rent-stabilized apartments empty?” and “good cause eviction in New York State covers Section 8 holders.”

Monto said that, as with surveillance, the city should implement a comprehensive policy regulating AI. Paniagua said she plans to have discussions with councilors and Owens on how to institute citywide standards for appropriate AI usage in official government capacities. 

“It’s totally worth exploring. And these are things that, if they’re gonna be used, then there’s got to be some measures,” Paniagagua said. “Because we can’t govern on AI.”

‘Dangerous territory’

When Johnson sent the AI materials to the council, mayor, and city lawyers, he disappointed Councilor Donna Moore, one of his allies in the fight to keep the Civilian Review Board operational.

Asked by a Central Current reporter about Johnson’s CRB legal brief purporting to speak in the voice of Thurgood Marshall, Moore — who has led the charge to preserve the CRB’s investigations into police misconduct complaints despite contrary legal advice from the city’s lawyers — said Johnson entered “dangerous territory” by disseminating those AI-generated materials.

Johnson in a Tuesday interview said that his background is in generative AI, prompt engineering and data analytics. He sent the documents as a “personal individual,” Johnson said, not as a representative of the board — though the email’s signature identifies Johnson as the chairman of the Citizen Review Board. 

“So I use AI almost exclusively in everything I do, and as a prompt engineer, I don’t get my models doing hallucination,” Johnson said. “… and I always have source support on any questions that or any document that’s generated through my AI prompting.”

Williams told Central Current that he dismissed Johnson’s missive after taking a cursory review of the documents and concluding that they were entirely generated by AI.

Sue Katzoff, the city’s lead counsel, in a responding email flagged to the other recipients on the email chain that her office had no role in the creation of the documents Johnson shared.

“Whomever drafted this document is misrepresenting their role and their client and I would ask that you immediately have this removed so no one is misled,” Katzoff wrote. “… Can you please provide further information and evidence that the document has been properly edited to reflect the attorney who drafted the submission and the client represented.”

Johnson said that he hadn’t consulted AI for help in deciding the next steps for the Citizen Review Board, which he chairs — but isn’t ruling it out, either.

“May. May not,” Johnson said. “But that’s how we go.”

Read more of Central Current’s coverage

Patrick McCarthy is a staff reporter at Central Current covering government and politics. A graduate of Syracuse University’s Maxwell and Newhouse Schools, McCarthy was born and raised in Syracuse and...

Leila Adelstein is a newsroom intern reporting on local news and writing features for the Central Current. She will be a junior at Hamilton College in the fall where she studies Public Policy and Studio...