disinformation

As a major storm rushed toward Florida last October, the head of the Federal Emergency Management Agency at the time faced a different kind of threat. Police had shown up in force to a rental property she owned as a result of a prank call, in a potentially dangerous attack known as “swatting.”

Back-to-back Hurricanes Helene and Milton had sparked a torrent of online conspiracies, with FEMA officials facing harassment and death threats, according to hundreds of pages of agency emails and other documents obtained via a Freedom of Information Act request by Bloomberg News. The records shed new light on how disaster-related misinformation affects the government’s emergency response, sucks up internal resources, and puts staff at risk.

Deanne Criswell, who ran FEMA under President Joe Biden, learned about the swatting situation as she was about to brief TV viewers on Milton, one of the most powerful storms on record to develop in the Gulf of Mexico. “It was a very unsettling feeling,” she said in a recent interview, thinking back on how she juggled her concern for her renters along with preparing Floridians for the storm.

FEMA Administrator Deanne Criswell testifies during a Senate Appropriations Committee hearing, November 20, 2024.

(Tom Williams/CQ-Roll Call, Inc via Getty Images)

Many of the attacks outlined in the documents have not previously been reported, including the doxxing of at least seven senior FEMA staffers. In those incidents sensitive personal information, such as home addresses, was published online for the purpose of harassment. The records also reveal challenges the agency faced as it tried to control the situation.

The incidents followed an online wave of disinformation suggesting FEMA was mishandling the response to the hurricanes that pummeled Florida and North Carolina in the lead up to the presidential election. Among the debunked claims swirling at the time were reports that agency workers had seized property from survivors and confiscated donations.

The offensive diverted agency time and resources to set the record straight and protect personnel. “It made my staff nervous,” said Criswell. “It made people in the community nervous. They didn’t know who to believe. They didn’t know who to trust.”The threat of misinformation continues to loom over the agency at a time when President Donald Trump and Department of Homeland Security Secretary Kristi Noem have made steep cuts to its staffing and funding, including pulling back on some of the resources FEMA used last fall to combat threats. In the aftermath of deadly Texas floods in July, for example, conspiracy theories online blamed cloud seeding.

“The profit-driven platform model, where sensational falsehoods outperform factual updates in emergencies, ensures this problem persists across political cycles and it can put lives at risk,” said Callum Hood, head of research at the nonprofit Center for Countering Digital Hate.

A FEMA spokesperson said in an email the agency “uses internal DHS resources to identify and mitigate any personal threats to employees.”

A trail of disinformation

Workers, community members, and business owners clean up debris in the aftermath of Hurricane Helene in Marshall, North Carolina, Sept. 30, 2024.

(Jabin Botsford/The Washington Post via Getty Im)

Hurricane Helene made landfall in the middle of the night on Sept. 26 as a Category 4 storm, causing historic flooding far inland and killing at least 250 people. Western North Carolina was particularly hard hit. Flood waters swept away small towns and cut off others, while Asheville lost water for more than a month. Almost immediately, FEMA staff had to confront false rumors circulating online, including that it had stopped accepting housing assistance applications from survivors and didn’t have enough funds to help them.

FEMA officials and experts attribute the quick spread of disinformation to historic government mistrust in the area, as well as social media platforms ratcheting back moderation. High-profile figures including X owner Elon Musk and Trump, then in the late stages of his bid to retake the White House, repeated some of the false claims. Trump, for example, said multiple times during his campaign rallies FEMA was directing disaster funds to immigrants.

For example, the agency shared a screenshot taken from a TruthSocial post from Oct. 5 that stated: “Deanne Criswell needs to be executed for crimes against humanity and treason!” An Oct. 6 post on Gab, a social media site favored by the far right, called for the “Mussolini treatment” of various officials. “The only question: Is there enough rope?” read one of the responses.

Jacyln Rothenberg, the agency’s spokesperson at the time, was among the most heavily targeted, leading Homeland Security to loan Customs and Border Protection agents to provide security at her home. “Because the doxxing was so severe and my safety was at risk, I had to stop tweeting,” she said. “I had to stop doing interviews. I had to stop putting myself on the record.”

FEMA staff also found what it called “far-right” users posting possible personal information for numerous officials, including Criswell, Coen and Rothenberg, internal documents show.

Attacks on FEMA Offline

As a second powerful hurricane — Milton — developed off the coast of Florida, the attacks on staffers’ started migrating from the internet to their homes. After Criswell’s rental property was swatted, among other “serious threats,” then-DHS Secretary Alejandro Mayorkas signed off on a government vehicle and extra security to protect the embattled FEMA chief.

Then it happened to someone else. “My deputy Jenna Peters’ home was swatted,” Coen told FEMA’s security team in an email on Oct. 11. Peters did not respond to a request for comment.

The most high-profile incident involved a man allegedly “hunting” FEMA staff in North Carolina’s disaster zone. On Criswell’s orders, she said in an email to other top Biden officials: “All FEMA staff and contractors working to interact with survivors and conducting housing inspections, as well as search and rescue teams stood down following the initial reports.”

Elena Gonzalez, 37, looks at their burned-out home after Hurricane Milton’s landfall on October 14, 2024, in Fort Myers, Florida.

(Eva Marie Uzcategui/The Washington Post via Getty Im)

Afterwards, FEMA put together a Workplace Protection Task Force involving security, intelligence and communications professionals to manage incoming threats. Protective measures included using specialized software to flag personnel previously targeted online as at risk of more harassment. But there were limits to how far the government could influence content moderation. At the time, outspoken Republicans led by House Judiciary Committee Chair Jim Jordan were investigating tech companies, alleging that the platforms were censoring conservative viewpoints under federal government pressure.

After initially approving ZeroFox to assist with facilitating takedowns, FEMA later asked that the company end all social media content removal requests. Per internal documents, the move came after staff discussions that it wasn’t advisable for the agency to contract for services that took any action beyond passive threat monitoring. ZeroFox declined to comment.

Supporters of 2024 Republican presidential candidate Donald Trump attend a boat parade near a house damaged in Hurricane Milton, Siesta Key, Florida, October 26, 2024.

(Chandan Khanna/AFP via Getty Images)

Trump’s team has already overseen a massive scale back of FEMA’s staffing, funding and programming. As part of a review of contracts, FEMA ended its agreement with ZeroFox, according to a former official familiar with the situation. A FEMA spokesperson confirmed that it ended the ZeroFox contract in April. For Melissa Ryan, founder of Card Strategies, a consulting firm that researches disinformation, the current political climate — in which public officials who attempt to provide transparency are often politicized and attacked — is a bigger obstacle than budget cuts in the fight against false claims. “So many of the new government appointees are Trump loyalists, and attempting to actually respond effectively to disinformation would make whoever made the attempt a target for MAGA and the administration,” she said.

Hirji, Alba and Leopold write for Bloomberg.

Manila, Philippines – When former Philippines President Rodrigo Duterte was arrested by the International Criminal Court (ICC) in March, Sheerah Escuerdo spoke to a local television station, welcoming the politician’s detention on charges of murder linked to his war on drugs.

Escuerdo, who lost her 18-year-old brother, Ephraim, to Duterte’s war, clutched a portrait of her sibling during the interview with News 5 Everywhere as she demanded justice for his killing.

Days later, she was shocked to find an AI-generated video of her slain brother circulating on Facebook, in which he said he was alive and accused his sister of lying.

“I’m alive, not dead. Are they paying you to do this?” the computer-generated image of Ephraim said.

The video, posted online by a pro-Duterte influencer with 11,000 followers, immediately drew thousands of views on Facebook.

One of the comments read, “Fake drug war victims”.

It was Escudero and her brother’s image from her News 5 Everywhere interview that the influencer had used to falsify their family’s tragedy. The video has since been reposted countless times, spreading to other social media platforms and resulting in Duterte supporters hounding Escuerdo daily.

“I wake up to hundreds of notifications and hate messages,” she told Al Jazeera.

“The worst thing is reading comments of people who believe this is real!” she added.

The same kind of harassment has been levelled at other vocal drug war victims, especially those under the group Rise Up, who actively campaigned for the ICC’s intervention.

Duterte’s arrest in March came amid a bitter power struggle between the ex-leader and his former ally, the incumbent president of the Philippines, Ferdinand Marcos Jr. Their alliance collapsed last year due to disagreements over policy, including Marcos Jr’s courting of the United States. The president’s supporters are now leading an effort to impeach Duterte’s daughter, Sara, from her post as the country’s vice president.

As tensions have escalated, supporters of Duterte and Marcos Jr have stepped up digital smear campaigns, using disinformation. Apart from fake accounts and doctored images, the disinformation mix has noticeably included AI-generated content.

Both the Marcos Jr and Duterte clans have been known to deploy disinformation tactics. Marcos Jr won the election in 2022 following a disinformation campaign that sought to whitewash his father Ferdinand Marcos’s brutal rule during the 70s and 80s.

But fact-checkers and experts say the recent uptick in posts peddling false narratives can be attributed more to the Duterte camp.

Disinformation nation

Victims of the drug war, their families, supporters and even their lawyers say incessant online disinformation has targeted them.

In a statement, the National Union of People’s Lawyers (NUPL), which represents Rise Up, a group of drug war victims, said the “online hate” was being “directed at widows, mothers, and daughters of drug war victims, attempting to intimidate them into silence”.

Both NUPL and Rise Up have now formally requested the government to investigate the increasing online harassment.

The campaign by Duterte’s supporters aims to discredit the ICC, demonise their detractors and paint their family as persecuted victims leading to and after the May 2025 mid-term polls, according to Danilo Arao, mass media expert and convener of election watchdog Kontra-Daya.

“The Duterte camp aims to deodorize the image of both patriarch and daughter. They will resort to disinformation to get what they want, even if it means twisting certain data,” Arao told Al Jazeera.

He pointed to posts circulating online that the ICC consented to grant Duterte’s request for an interim release, which in reality was denied.

The surge in disinformation has caused worry among Filipinos.

A report released in June by Reuters Digital News found that a record number of Filipinos – nearly 7 out of 10 – were more concerned with disinformation than ever before.

In the same month, Duterte-allied senator, Ronald Dela Rosa, shared an AI-generated video on his official Facebook page. The video, which showed a young man criticising the “selective justice” targeting Sara Duterte, was posted on June 14, garnering at least 8.6 million views before it was taken down.

The vice president defended the video, saying there’s “no problem sharing an AI video supporting me as long as it’s not for profit”.

Arao, the mass media expert, countered, saying the politician is trying to normalise disinformation, and that she “badly needs media literacy”.

Tsek.ph, the Philippines’ pioneer fact-checking coalition, noted that fact checks on posts about Duterte’s ICC arrest over a six-week-period account for almost a quarter of the 127 news articles curated by the group.

The figure surpasses the two dozen pieces of news related to Sara Duterte’s impeachment.

On Sara Duterte’s deepfake defence, Tsek.ph coordinator Professor Rachel Khan told Al Jazeera that “for the educated, it reinforces their already tainted image of disregarding truth. But for followers, it could reinforce the dictum that ‘perception is truth.’”

In reality, the popularity of the Duterte family has waned significantly.

Opinion and approval surveys conducted in March indicate that at least 51 percent of the public want Rodrigo Duterte to be tried for his alleged crimes. Likewise, polls in June found that at least 66 percent of people want Sara Duterte to confront allegations of corruption against her through an impeachment process.

AI growth

The government of last year launched a task force to mitigate disinformation and the use of AI. However, spikes in disinformation were already noticeable in December as the Marcos-Duterte rivalry heated up.

Tsek.ph tracked the increasing use of AI in disinformation before the mid-term elections held in May this year. It found that from February to May, out of 35 unique altered claims, nearly a third “likely involved deepfake technology to impersonate public figures or distort reality”.

“This is a problem of human behaviour, not AI. It’s a disinformation influence operations problem, exacerbated by the unethical usage of AI tools,” Carljoe Javier, executive director of Data and AI Ethics PH, told Al Jazeera.

All mainstream political forces in the Philippines have, to some extent, deployed AI technologies to boost their agendas. The latest OpenAI Safety Report revealed that Comm&Sense, a Manila-based tech firm, used AI for a campaign using thousands of pro‑Marcos Jr and anti‑Duterte comments across Facebook and TikTok.

Besides generating content, the firm also used AI to analyse political trends and even draft public relations strategies.

The report said Comm&Sense manufactured TikTok channels to post identical videos with variant captions while handling shell accounts to post comments and boost engagement.

The use of AI to outline plans, not just create content, marks a shift away from the Marcos Jr administration employing troll armies as he did in his 2022 campaign.

“If you have the resources and the bully pulpit of the government, you can afford to keep on swatting the Dutertes and their partisans for whatever statements they have made against the Marcos government,” said Joel Ariate Jr, a researcher tracking political developments at the University of the Philippines Third World Studies Center.

“If you put AI in the hands of an already good public relations or marketing team, the capacity for disinformation is amplified by so much. They can have one message and instantly generate 20 different versions of it,” explained Javier.

The Philippines has several pieces of legislation in congress concerning the responsible use of AI. For a healthy policy approach, Javier believes that technical and ethical experts would be crucial.

He said he hoped the country’s leaders can take important steps, but said he has doubts about their appetite for ethical AI legislation.

“Is there enough push for legislators to advance a policy given that they may be benefitting from the current state of political operations?” he asked.

Published on
24/06/2025 – 18:17 GMT+2

ADVERTISEMENT

The Brazilian ambassador to the EU has told MEPs in Brussels that a disinformation campaign surrounds the trade deal signed in December 2024 between the EU and the Mercosur countries – Argentina, Brazil, Paraguay, and Uruguay.

Pedro Miguel da Costa e Silva strenuously countered the arguments of the deal’s critics during a hearing of the Parliament’s trade committee on Tuesday.

“The occurrence of animal diseases is much higher in the EU than in Brazil. It shows the need to check the veracity of some narratives,” the ambassador said, holding up a sheet of paper and adding: “In any case, I need to stress that nothing in the agreement changes the right of the EU and its member states to protect human, animal, or plant health.”

The Mercosur agreement aims to establish a transatlantic free trade zone encompassing 750 million people and nearly one-fifth of the global economy.

The EU member states have yet to adopt the deal, but some – led by France – oppose it, facing strong domestic resistance from environmental activists and farmers who argue that it would create unfair competition and fail to uphold environmental and phytosanitary standards.

“The debate about this agreement has not always been a balanced one. Some people want to apply a unique benchmark to Mercosur and ask us to engage in an endless loop of negotiations,” Da Costa E Silva said.

He denounced what he described as unfair treatment of the deal when compared to others the EU has negotiated – citing recent agreements between the EU with Chile or Mexico, and those under discussion with India and the US – claiming these haven’t faced the same kinds of “accusations, and unreasonable demands and expectations”.

The ambassador also sought to counter the arguments raised by farmers concerned that their Brazilian counterparts would gain unfair competitive advantages.

“The [market] access we received in products considered sensitive by the European producers is very limited,” he said. And he claimed that some Brazilian standards are more stringent than European. “For example: the share of land that our farmers need to set aside for the protection of native vegetation varies from 20% of their properties in the south of Brazil to 80% in the Amazon region. This is far beyond the requirements asked of European farmers.”