Somewhere inside Iran, a network of activists exchanged messages about politics and news with an activist — nicknamed “Parsa” — on the messaging app Telegram. They had no idea they were actually speaking to an AI agent that had copied the real activist’s style by training it on 8,400 real Telegram posts.
Instead of “Parsa,” they were actually communicating with the People’s Mojahedin Organization of Iran (MEK), an Iranian Islamist-Marxist opposition group widely regarded as a cult.
According to a report from Anthropic released last month, this operation was carried out by four individuals working for the MEK. They built an AI agent named “Viktor” that also surveilled individuals inside Iran and flooded social media platforms with pro-MEK content under “independent” branding.
The Anthropic report sheds light on how foreign actors are increasingly using artificial intelligence tools like Claude and ChatGPT for their influence operations. AI agents like Viktor are software systems powered by large-language models that can operate autonomously. A lot of the actions described in the report — like drafting reports, creating content, compiling information — looks like ordinary work similar to what millions of Americans use Claude and ChatGPT for.
Perhaps the report’s biggest surprise is the degree to which U.S.-aligned groups and partners have embraced AI for their influence operations.
The MEK, for example, boasts notable support on Capitol Hill. Last year, 230 members of the House co-sponsored a resolution urging U.S. support for the MEK’s plan for Iran. Former National Security Adviser John Bolton, former Vice President Mike Pence, and former New York City Mayor Rudy Giuliani are just a handful of high-profile Republican politicians who have spoken at MEK rallies, often for a hefty speaking fee.
One account managed by the MEK-linked “Viktor,” @Tehranchekhabar19, boasts over 170,000 followers on Instagram. The account presents itself as an independent news account while posting content that support the MEK’s views, such as videos that promote the U.S.-Israeli war against Iran while attacking Reza Pahlavi, the former Crown Prince of Iran who leads a rival opposition faction. The account’s videos sometimes garner hundreds of thousands of views. “Viktor” also set up other Instagram accounts boasting hundreds of thousands of followers, some of which appear to have since been shut down.
The MEK, which the U.S. listed as a Foreign Terrorist Organization until 2012, was even more active on Telegram to try to covertly recruit people inside Iran to support the MEK’s agenda. According to Anthropic, “Viktor” infiltrated more than 500 groups on Telegram, then “grouped these targets by city, age, occupation, political alignment, and arrest history likely to help them tailor their messages to the specific audiences.”
Craig Holman, Government Affairs Lobbyist for Public Citizen, explained in an email to RS that this marks a new type of influence tactic. “These accounts of foreign governments using AI to manipulate the opinions of legislative bodies and the public are based on the root source of AI deception: falsely depicting a real person or entity doing or saying something that they never did.”
This case study and others described in similar reports reveal that foreign actors are taking concrete steps to add AI to their arsenal of weapons to win information wars. For instance, a report from OpenAI published last week detailed how Iranian operators successfully placed almost 100 articles in different international affairs-focused publications under phony bylines using ChatGPT. The Iran-written articles typically promoted an anti-Iran War point of view, with headlines such as “Is America’s Iran Policy Being Privatized? The Jared Kushner Question” and “Trump’s Iran Gamble Hardened the Regime It Meant to Break.”
The recent OpenAI report also found that Russia had successfully created and controlled a fake think tank called the “Social Research Center” with real staff on the ground in Latin America who had no idea they were working for Russia. Russian operators used ChatGPT to create content and draft internal reports. Meanwhile the unknowing staff carried out real interviews with experts and published research papers, mostly about India-Latin America relations.
While the Social Research Center’s X account has been suspended, the organization’s website is still active as of publication. There, it posted a statement rejecting OpenAI’s claims as “factually incorrect and as constituting a deliberate attempt to discredit an independent research project.” Another case study revealed how one of Washington’s closest partners in the Middle East is embracing AI to absolve itself of its role in a bloody civil war.
Emirati officials created their own AI agent on Claude, nicknamed “Deadshot” — and charged it with the herculean task of whitewashing the UAE’s role in the Sudanese civil war. A UAE-backed Sudanese paramilitary group, the Rapid Support Forces (RSF), has been crediblyaccused of ethnic cleansing and sexual violence. A U.N. fact-finding mission had said that the RSF’s actions against non-Arab communities in Sudan’s Darfur region bear the “hallmarks of genocide.”
The AI agent got right to work to maintain the UAE’s reputation, targeting the U.N. in particular. According to the Anthropic report, “Deadshot” hijacked the identity of a real Sudanese human rights organization and ghost-wrote U.N. testimony in hopes that the world body would be influenced by “independent local witnesses, rather than state messaging.”
Anthropic’s report gives conflicting accounts as to whether the effort was successful. At one point, it says it could not confirm whether the testimony reached its intended audience, though a summary elsewhere in the report describes it as having been “delivered in a live UN Human Rights Council session.” Anthropic did not respond to a request for comment.
Deadshot compiled information on the UAE’s critics, creating “counter-accountability dossiers” on U.N. Special Rapporteurs who had criticized the UAE’s role in Sudan.
The UAE also carried out a broader information campaign centered primarily on X. “Deadshot” created and managed a network of around 300 inauthentic social media accounts that posted content to pin the blame for the Sudanese civil war on the Muslim Brotherhood, which is supporting the Sudanese Armed Forces against the RSF.
The posts themselves often make humanitarian appeals, deflecting blame for the war away from the RSF and toward the Sudanese Muslim Brotherhood, which the U.S. designated a Foreign Terrorist Organization earlier this year. “The longer Sudan’s war drags on, the deeper the famine crisis becomes. Analysts point to hardline Islamist elements supporting continued military confrontation over negotiated settlements,” reads one post on X by a user with 27,000 followers that Anthropic says was part of the coordinated campaign. Another post on X from a user with 119,000 followers says the SMB role in the conflict “makes ending the war far harder.”
According to the report, the operator instructed Claude to carry out a “coordinated transatlantic and regional operation to dismantle the Muslim Brotherhood globally” hundreds of times.
Kholood Khair, Founding Director of Khartoum-based think tank Confluence Advisory, explained in an interview with RS that the Sudanese conflict is being fought on three fronts: the battlefield, the mediation table, and the information space — and that last one is becoming more crucial. “These posts are carefully deployed to attempt to co-opt the pro-civilian rule, pro-democracy movement and camouflage what the UAE is actually trying to achieve in Sudan,” said Khair. She noted that, while many of the posts aren’t factually incorrect, they are selective in the narrative they tell.
“AI is the next battleground in many of these wars, and Sudan’s war has clearly become a testing ground, with the UAE getting a head start, ” Khair added.