News
WASHINGTON, DC, USA –– White House Press Secretary Karoline Leavitt answers questions from journalists on Jan. 28, 2025.
WASHINGTON — As the Trump administration increasingly welcomes vloggers and social media influencers into press briefings and the Oval Office, established outlets like the Associated Press find themselves excluded from the century-old press pool, sparking controversy about what "transparency" truly means.
Joshua Sukoff is a photojournalist from Long Island, New York. He is currently studying journalism at Northwestern University.
A 2026 study found AI beat elite debaters, until it was slowed to human speed. Why persuasion throughput matters and how to make chatbot tempo transparent.
A political chatbot can answer before a citizen has finished reading its previous message. It can add statistics, historical examples, causal claims, and links at a rate no human campaign volunteer could match. That does not necessarily make its case wiser. It changes the contest by allowing one speaker to control the tempo.
A June 2026 preprint offers unusually direct evidence of this problem. Across four preregistered experiments involving 18,978 conversations from 6,923 people, frontier AI systems were more persuasive than laypeople, tournament winners, professional canvassers and elite competitive debaters. The human experts were not casual opponents: the study included world and continental champions, paid preparation, live practice and substantial performance incentives.
The most revealing result came when the researchers changed the rules of the exchange. Elite debaters had produced replies averaging 54 words after roughly 95 seconds. The AI produced about 294 words with sub-second latency. When the system was constrained to roughly human-length messages and human writing speed, its advantage over coached elite debaters fell to a statistically indistinguishable zero.
The study is a preprint, not a final verdict, and its experiments were conducted in controlled text conversations in the United Kingdom. It does not prove that any chatbot will decide an election. It does, however, identify a democratic variable that is largely missing from public debate: persuasion throughput, or the amount of argumentative material a system can deliver within a citizen's limited attention.
Our current safeguards focus on important questions. Was the message generated by a machine? Who paid for it? Was personal data used to target the recipient? Is the image or voice synthetic? Those questions address identity and targeting, but they do not address what happens after a person enters the conversation.
A chatbot can present dozens of claims before the user has inspected the first source. In the 2026 study, unconstrained AI produced about 37 fact-checkable claims per conversation; the human-paced version produced about 12. Across conditions, fact density strongly predicted persuasive impact. The advantage was not simply that the machine sounded more empathetic or more human. It could place more information on the table before its counterpart could answer.
A peer-reviewed Science article published in December 2025 points in the same direction from another angle. In experiments with 76,977 participants, 19 models and 707 political issues, post-training and prompting designed for persuasion mattered more than personalization or model scale. The same interventions that increased persuasion also reduced factual accuracy. Speed and density can therefore amplify useful information and error at the same time.
The appropriate response is not a government word limit on political speech. It is to make the tempo of machine persuasion visible and controllable. A political chatbot should display how many new factual claims it has introduced, how quickly it is producing them, and whether the user has asked for additional material. A citizen should be able to switch to a slow mode that presents one claim at a time, pauses until the evidence has been opened, and offers a neutral summary before moving on.
Sources should be attached to claims, not scattered as decorative links at the end of a long answer. The interface should separate established facts, contested interpretations, and predictions. It should also let the user request the strongest counterargument without forcing the person to leave the conversation and find an opposing system.
Political chatbots should identify the model version, the date of the governing instructions, and the organization responsible for the conversation. Campaigns, parties, advocacy groups and platforms should preserve sample transcripts for independent audits. Those audits should measure claim accuracy, correction behavior, source diversity, and whether the system changes its standards when arguing for different sides.
These measures are viewpoint-neutral. They do not decide which position is correct or prevent a campaign from making a forceful case. They address a procedural asymmetry: one participant can generate arguments at machine speed while the other remains bound by human reading, memory, and verification.
Disclosure alone will not solve that asymmetry. A label saying that the speaker is a bot tells the citizen who is talking, but not how the interface is shaping the exchange. A clearly identified system can still overwhelm the user with a sequence of claims that arrives too quickly to compare, challenge, or remember. Transparency must therefore cover the mechanics of persuasion, not merely the identity of the persuader.
A useful standard could be tested without deciding which political arguments are permissible. Does the system count and display new factual claims? Can the user stop automatic continuation? Are sources attached to the precise claims they support? Can an independent reviewer reproduce a sample conversation under the recorded model and instructions? These are measurable design questions, and campaigns could be compared on them before an election rather than after a controversy.
The limits of the research matter. The conversations lasted a median of about 14 minutes, participants were paid, and the tested political issues were British. Real-world influence will depend on exposure, trust, repetition, and whether people willingly sustain political conversations with a machine. That uncertainty is a reason to test safeguards, not a reason to ignore the mechanism.
Democratic deliberation requires more than access to an answer. It requires time to understand, compare, and contest the answer. We do not need to ban machine speech to protect that space. We need to stop treating the machine's ability to set the pace as if it were a neutral feature of the interface.
A democracy in which one side can generate an argument faster than the other side can read it is not necessarily better informed. It may simply be losing control of the clock.
Roney Lima do Nascimento is a mathematics educator, AI specialist and doctoral candidate in Pure Mathematics at the University of São Paulo. He writes about model evaluation, education, democracy and institutional capacity. His work has appeared in Folha de S.Paulo, Nexo Políticas Públicas, Congresso em Foco, EUobserver, GovInsider, HEPI and Educational Leadership.
U.S. President Donald Trump is seen in the electronic viewfinder of a television camera in the Oval Office at the White House on September 18, 2026, in Washington, DC.
The Fulcrum is committed to nurturing the next generation of journalists.
We asked Malory Shaw, a Fall Fulcrum Fellow, to share what democracy means to her and how she sees its current health.
Local news informs voters, and informed voters fuel democracy. It holds politicians accountable, even when they go unnoticed, and it helps voters choose the candidates who will serve them best.
I experience democracy and a free press firsthand in the public media newsroom where I intern. We interview every candidate and ask about the issues voters care about, from police surveillance to data centers in northern Nevada's deserts.
I interviewed Reno’s new police chief as he settled into his role, appointed by the City of Reno. Our broadcast let listeners judge for themselves whether the decision to appoint him was right, given concerns about surveillance and immigration. We keep a record of what officials promise, so voters can look back and see: what was fulfilled and what wasn’t?
Local reporting connects with people like nationals may not. If our station or other local outlets don’t cover it, no one else will.
The free press has deteriorated with democracy. In the U.S., democracy has exponentially declined. The International Institute for Democracy and Electoral Assistance's annual Global State of Democracy report reveals that the U.S. has reported an all-time low in democracy under President Donald Trump’s administration. I don’t believe we’ve ever had a robust, healthy democracy where every group was represented. But the decline is sharper under President Trump, whose attacks on the press are attacks on democracy itself. Democracy is the basis of our country.
President Trump has banned CNN, Politico, and MS NOW, citing false and inaccurate reporting. On Sept. 21, the White House launched Trump TV, which livestreams government highlights 24/7. These attacks aren’t new under his administration. The Pentagon required journalists to sign off on restrictions, such as having an escort and reporting on news with official approval.
These restrictions control the narrative and create gaps in what the public can know. In some Latin American countries, state-run media only highlights the government’s successes. Democracy is based on knowledge, and without it, how do we know how to vote? That is feigned democracy, in name only but not in practice.
In a functional democracy, the press should not be restricted in any form or state vetting.
Brazen restrictions aren’t the only tactic. Journalists at protests are pepper-sprayed and mistreated, which discourages others from doing the work. A chilling effect the First Amendment is meant to dissuade.
If journalists are brutalized and expected to be harsh, nobody would shame them for feeling dissuaded. Nobody should blame a reporter for feeling deterred. The most passionate and committed journalists keep working despite this, but there is a limit to what a reporter should endure to get information to the public.
The University College London Social & Historical Sciences department says maintaining a healthy democracy requires politicians to avoid eroding public trust in the media. President Trump releases barrages of insults against the media he dislikes, helping create a polarized country that is unable to trust the mass media.
No media should be trusted blindly, but its role in democracy remains and should be treated with incredible respect.
Public funding and support is not a favor to the journalist. It is securing an investment in a healthy democracy where accountability is at the forefront.
Politicians must go the extra step and advocate for public media funding. The newsroom I intern with holds politicians accountable with minimal resources. Our station has four impassioned reporters and several interns, but how can a station this size fully cover Reno and all of northern Nevada? More reporters would mean a stronger democracy here.
Governments and politicians have a duty to give reporters the resources they need to maintain a healthy democracy.
Malory Shaw is a student at the University of Nevada, Reno. Malory's reporting is grounded in building trust with communities — from pesticide‑affected families in rural Costa Rica to Latino neighborhoods in Nevada — and in examining how law enforcement technology impacts democracy.
Influencer and left-wing political commentator Hasan Piker is introduced during a campaign rally for Wisconsin Gubernatorial candidate Francesca Hong on Aug. 2, 2026, in Milwaukee.
No one likes to think they’re being hysterical. And nothing makes hysterical people angrier than being described as such. So, let me offer my apologies in advance for what follows.
Jeremy Musighi, an independent researcher, recently released a study analyzing 14.7 million news articles published across 65 languages over the last decade. He found that, in 2024, news outlets published “more than twice as many stories about Israel as about all 48 countries of sub-Saharan Africa combined.”
Given that many countries focus on Israel to distract from their own problems, that’s not that shocking. But Musighi also found that, “On American cable news, Israel was the most-mentioned country— ahead of Iran, Russia, Ukraine, China. On the BBC’s news channel, 1 in every 17 broadcast minutes mentioned Israel.”
One might argue that given Israel’s recent actions, it earned such coverage. But Israel gets wildly disproportionate coverage in peacetime too. In 2019, long before the war in Gaza, “1 of every 112 news articles on Earth — from all the world’s outlets minus Israel’s — was primarily about Israel. 1.8 million international articles mentioned it.”
Now, I have reservations about Musighi’s methodology, given how media outlets recycle and syndicate content. But rather than wade into the statistics, for argument’s sake let’s just cut all of his findings in half. That would still be hard to defend rationally.
Also, this says nothing of the quality of the reporting. If it was objective, the quantity wouldn’t matter much. But a sizable portion is objectively anti-Israel, often taking Hamas’ claims at face value. When Hamas shrieked that Gaza was on the brink of famine, the Western press breathlessly repeated the claim, which was utterly false.
In a recent interview on Taqarrab, an Arab-language podcast, acclaimed photo “journalist” Motaz Azaiza admitted that he refused to publish images of Hamas abuses — to protect the abusers. The New York Times’ Nick Kristoff launched weeks of coverage with the claim that Israelis train dogs to rape Palestinians. American media covered the controversy over Macklemore’s criticisms of Israel like a free speech martyr.
In fairness, a lot of coverage is driven by news, like when the United Nations condemns Israel, which it has done 187 times since 2015 — nearly six times more often than it has condemned Russia, and 17 times more often than Iran or North Korea.
Founded in 1948, modern Israel is middle-aged as countries go.
Look at a list of 195 countries in order of their founding date and Israel ranks somewhere between the 85th- and 90th-oldest (if we acknowledge biblical Israel, then it’s the fourth-oldest nation in the world). And yet many talk about Israel as if it’s a sort of recent mistake, easily remedied, like outgoing mail that hasn’t been picked up yet.
Israel is only slightly older than modern Tibet, which was, according to the Chinese government, “liberated” in 1951. Many Tibetans use the term “occupied.”
We don’t hear much about “occupied Tibet” from people fixated on the evils of occupation. Nor do we hear about China’s system of apartheid— another charge casually hurled at Israel. Nor do we hear about China’s genocidal policies toward the Muslim minority Uyghur population in Xinjiang. The charge of cultural genocide is irrefutable, but there’s also strong evidence of more literal genocide thanks to the regime’s violent and coercive reproduction policies, slashing birthrates in half.
Nor do we hear much about China’s policy of settler-colonialism in Tibet and Xinjiang, where Han Chinese are incentivized to migrate as part of the regime’s policy of “sinicization.” I think China’s pretty powerful, but criticizing Israel, according to the popular online streamer Hasan Piker (who will not condemn China), is “speaking truth to power.”
To that end, we hear a lot about Israel’s occupation, settler-colonization and genocide. We don’t have the space to contend with all of these charges, so for the sake of brevity and reasonableness let’s concede, for now, the first two. Besides, it’s the false genocide charge that fuels anti-Israel obsessions more than any other. The claim is more than 40 years old. And yet, since the founding of Israel, the Palestinian population has grown more than fivefold.
When you hear that Israeli genocide proves Israelis are the “new Nazis” keep in mind that the Holocaust killed 2 out of every 3 European Jews.
A new documentary, “Naza,” that purports to prove that Israel is committing genocide bases its claim on the fact that Israel goes to great lengths to track collateral damage. Never mind that there’s no such thing as collateral damage when you intend genocide.
It should take the average reader five minutes to read this column. When “Naza” debuted at the Venice Film Festival this month, the audience’s standing ovation lasted five times as long. That, I’m sorry to tell you, is a sign of hysteria.
Jonah Goldberg is editor-in-chief of The Dispatch and the host of The Remnant podcast. His Twitter handle is @JonahDispatch.
New research shows AI beats elite debaters in persuasion — not through better arguments, but by out-pacing human reading and reply speed.
A political chatbot can answer before a citizen has finished reading its previous message. It can add statistics, historical examples, causal claims and links at a rate no human campaign volunteer could match. That does not necessarily make its case wiser. It changes the contest by allowing one speaker to control the tempo.
A June 2026 preprint offers unusually direct evidence of this problem. Across four preregistered experiments involving 18,978 conversations from 6,923 people, frontier AI systems were more persuasive than laypeople, tournament winners, professional canvassers and elite competitive debaters. The human experts were not casual opponents: the study included world and continental champions, paid preparation, live practice and substantial performance incentives.
The most revealing result came when the researchers changed the rules of the exchange. Elite debaters had produced replies averaging 54 words after roughly 95 seconds. The AI produced about 294 words with sub-second latency. When the system was constrained to roughly human-length messages and human writing speed, its advantage over coached elite debaters fell to a statistically indistinguishable zero.
The study is a preprint, not a final verdict, and its experiments were conducted in controlled text conversations in the United Kingdom. It does not prove that any chatbot will decide an election. It does, however, identify a democratic variable that is largely missing from public debate: persuasion throughput, or the amount of argumentative material a system can deliver within a citizen's limited attention.
Our current safeguards focus on important questions. Was the message generated by a machine? Who paid for it? Was personal data used to target the recipient? Is the image or voice synthetic? Those questions address identity and targeting, but they do not address what happens after a person enters the conversation.
A chatbot can present dozens of claims before the user has inspected the first source. In the 2026 study, unconstrained AI produced about 37 fact-checkable claims per conversation; the human-paced version produced about 12. Across conditions, fact density strongly predicted persuasive impact. The advantage was not simply that the machine sounded more empathetic or more human. It could place more information on the table before its counterpart could answer.
A peer-reviewed Science article published in December 2025 points in the same direction from another angle. In experiments with 76,977 participants, 19 models and 707 political issues, post-training and prompting designed for persuasion mattered more than personalization or model scale. The same interventions that increased persuasion also reduced factual accuracy. Speed and density can therefore amplify useful information and error at the same time.
The appropriate response is not a government word limit on political speech. It is to make the tempo of machine persuasion visible and controllable. A political chatbot should display how many new factual claims it has introduced, how quickly it is producing them, and whether the user has asked for additional material. A citizen should be able to switch to a slow mode that presents one claim at a time, pauses until the evidence has been opened, and offers a neutral summary before moving on.
Sources should be attached to claims, not scattered as decorative links at the end of a long answer. The interface should separate established facts, contested interpretations, and predictions. It should also let the user request the strongest counterargument without forcing the person to leave the conversation and find an opposing system.
Political chatbots should identify the model version, the date of the governing instructions, and the organization responsible for the conversation. Campaigns, parties, advocacy groups, and platforms should preserve sample transcripts for independent audits. Those audits should measure claim accuracy, correction behavior, source diversity, and whether the system changes its standards when arguing for different sides.
These measures are viewpoint-neutral. They do not decide which position is correct or prevent a campaign from making a forceful case. They address a procedural asymmetry: one participant can generate arguments at machine speed while the other remains bound by human reading, memory, and verification.
Disclosure alone will not solve that asymmetry. A label saying that the speaker is a bot tells the citizen who is talking, but not how the interface is shaping the exchange. A clearly identified system can still overwhelm the user with a sequence of claims that arrives too quickly to compare, challenge, or remember. Transparency must therefore cover the mechanics of persuasion, not merely the identity of the persuader.
A useful standard could be tested without deciding which political arguments are permissible. Does the system count and display new factual claims? Can the user stop automatic continuation? Are sources attached to the precise claims they support? Can an independent reviewer reproduce a sample conversation under the recorded model and instructions? These are measurable design questions, and campaigns could be compared on them before an election rather than after a controversy.
The limits of the research matter. The conversations lasted a median of about 14 minutes, participants were paid, and the tested political issues were British. Real-world influence will depend on exposure, trust, repetition and whether people willingly sustain political conversations with a machine. That uncertainty is a reason to test safeguards, not a reason to ignore the mechanism.
Democratic deliberation requires more than access to an answer. It requires time to understand, compare and contest the answer. We do not need to ban machine speech to protect that space. We need to stop treating the machine's ability to set the pace as if it were a neutral feature of the interface.
A democracy in which one side can generate an argument faster than the other side can read it is not necessarily better informed. It may simply be losing control of the clock.
Roney Lima do Nascimento is a mathematics educator, AI specialist and doctoral candidate in Pure Mathematics at the University of São Paulo. He writes about model evaluation, education, democracy and institutional capacity. His work has appeared in Folha de S.Paulo, Nexo Políticas Públicas, Congresso em Foco, EUobserver, GovInsider, HEPI and Educational Leadership.
