In the Media

WNYC and Gothamist reporter Karen Yi meets Carolina Abreu, whose pitch-black Bronx street got streetlights two days after her video racked up nearly 2 million views. Then NYU politics professor Joshua Tucker explains what’s behind the trend and where the limits are when a request goes from fixing a pothole to changing a policy.

Jennifer Allen, a professor at NYU Stern who worked at Facebook from 2016 to 2018, has researched the effect of misinformation and vaccine-skeptical content on Facebook. She and her colleagues have estimated that exposure to vaccine-skeptical content on the platform during the Covid-19 pandemic reduced people’s intention to get vaccinated. Without that exposure, they estimated 5 million more Americans would have been vaccinated, saving approximately 21,000 lives.

Even if misinformation doesn’t make up the majority of most people’s media diet, Nagler says, it can be weaponized to speed up bad actors’ misinformation campaigns and translate them for a wider online audience. “We would be foolish not to worry about AI influencing this.”

Documented worked with the NYU Center for Social Media, AI, and Politics (CSMAP) to access an exclusive collection of tens of millions of TikTok video URLs from a large-scale representative sample of TikTok posts published since October 2024. With access to this database, Documented was able to extract videos from a 20-month period — from November 2024 to July 2026 — and build a sample of roughly 10,000 of the most-watched videos related to immigrant issues in New York City.

Social media, whatever its flaws, mostly happened in public, noted Joshua Tucker of New York University, who co-leads the American Political Science Association’s AI task force with Persily. “Where we’re going to have an enormous blind spot is actually how people are using these models, because the vast, vast, vast majority of interactions that people have with chatbots are in one-on-one circumstances, which is an essentially private conversation,” he said.

“I’m generally not opposed to the sort of labels that exist, on narrowly targeted and materially deceptive content,” said Scott Babwah Brennen, the director of New York University’s Center on Technology Policy and a policy research affiliate at its Center for Social Media, AI, and Politics.

A study published in 2023 by N.Y.U.’s Center for Social Media and Politics, in Nature Communications, found that Russian bots on Twitter had limited influence, with one per cent of users being exposed to seventy percent of the posts created by Russian troll accounts.

“The idea in a democracy is that voters are presented with factually correct information about candidates, and then they make their choices," Tucker said. "And any time you have people who are trying to obscure those factually correct positions of candidates or actually correct things about candidates, that is, in a sense, undermining some of the normative basis of democracy.”

“People often talk about AI as if it learns from the internet in some neutral way. It doesn’t,” said Hannah Waight. “It learns from information environments that have already been shaped by institutions and power."

"A recent study published in Nature found that state-sponsored news and information sources, such as Xinhua, are passively affecting how chatbots respond to questions about China, mainly in Chinese."

"In a new paper, Hannah Waight of the University of Oregon and her co-authors put politically charged questions in English and 37 other languages to OpenAI’s GPT-3.5 and other models. “State control of media affects language model outputs through its appearance in training data,” the authors conclude."

Joshua Tucker, one of the authors of the study, explains in an interview how governments exert indirect influence - and where this could lead in the future.

"Last week, while the world watched President Trump and Xi Jinping meet in Beijing, a much quieter piece of research dropped in Nature. It should have been on every front page."

"Researchers from several U.S. universities—including the University of Oregon, New York University, and Princeton University—argue in the journal Nature that governments may indirectly shape large language models by influencing the online information used to train them." (translated from German)

Joshua Tucker joins 'Squawk Box' to discuss details of the university’s new research showing that AI chatbots generate more government-friendly responses when prompted in languages tied to countries with tighter media controls, the intersection of politics and AI, and more.

Multiple co-authors from our recent paper, "State media control influences large language models," spoke with India's The Week on how content from state-controlled media can influence information environments.

By using the same language, Melina Much said, Miller keeps the phrase high-level and recognizable to the general public while reframing it with new connotations. It’s part of a broader form of cultural in-group signaling.

Even as AI tools transform misinformation, says Jennifer Allen, “I think that they have [also] transformed the field.”

Joshua Tucker speaks at the University of Konstanz, showing that government control of media influences LLM outputs through training data, using results from China and a cross-national audit.

Melina Much discusses political officials’ use of podcasts with the Associated Press, given podcasts’ “massive” reach when videos are clipped and shared across social media platforms, and their more intimate, conversational, and friendly setting, which allows administration officials to promote themselves without facing as much pushback.

Joshua Tucker joins ‘Squawk Box’ to discuss the landmark social media addiction case against Meta and YouTube, how social media has evolved over the years, and more.

Joshua Tucker commented on Zohran Mamdani's campaign strategy, which prioritized algorithmic engagement through high-quality videos explaining policies.

Melina Much speaks with American Friction hosts on the politics and impacts within popular podcasts.

Benjamin Guinaudeau speaks with NPR concerning recent allegations of TikTok censorship.

Currying political capital online involves demonstrating that the opinions you hold are widely held while also engineering viral content that aligns with your views, says Joshua Tucker.