The 2023 Maui wildfires killed more than 100 people, destroyed nearly 2,000 homes and nearly erased a historic island community. The fires began after a broken power pole owned by the local electric utility sparked against dry vegetation during a windy day.

Yet in the aftermath, thousands of people across major social media platforms shared posts and videos claiming that the disaster had a more sinister cause. The posts alleged that the government had used a “direct energy weapon,” or laser, to ignite the fires. Depending on the version, the supposed attack was part of a plan to reclaim the area for a high-speed rail line or develop a high-tech “smart city.”

Right-wing activists and social media personalities, including the late Charlie Kirk, amplified some of the claims. The Federal Emergency Management Agency also observed accounts linked to China circulating rumors that a “weather weapon” had caused the fires. Other accounts claimed FEMA would seize the homes of people in Maui who applied for emergency aid, reviving traumatic memories of colonization in a community with significant Indigenous historical and cultural ties.

This article is part of Distorted, a Grist series examining misinformation after extreme weather events. It was published with support from the CO2 Foundation.

A worsening information crisis

Rumors have always spread during disasters, but the conditions for misinformation have become more severe. Climate change is producing supercharged wildfires, rapidly intensifying hurricanes, devastating droughts and sudden floods that leave communities confused and disorganized. At the same time, the information environment has become increasingly fragmented.

Major social media companies have abandoned much of the pretense of comprehensive fact-checking and content moderation. Politicians frequently distort or politicize disasters. Government agencies are short-staffed following DOGE cuts. Generative artificial-intelligence systems have created a new and poorly understood way for people to obtain information.

The problem extends beyond isolated social media users. A recent News Literacy Project report found that a vast majority of teenagers encounter conspiracy theories on social media, and that a vast majority of teens who encounter those theories are inclined to believe them.

“Disasters are, by their nature, very urgent and have a lot of uncertainty surrounding them,” said Amber Silver, a professor of emergency management at the University at Albany who specializes in disaster communications. People affected by extreme weather urgently need facts, she said, but often have almost no reliable way to obtain information about their situation.

“So now we have essentially all of the perfect conditions for the spread of misinformation,” Silver said.

The consequences extend beyond confusion. An information vacuum after a disaster can cost lives, contribute to traumatic displacement and slow recovery. It can also damage the relationship between public authorities and the people they serve.

Research from around the world illustrates the effects. A Department of Homeland Security study found that an inaccurate map posted on social media during a 2017 California flood caused an unnecessary mass evacuation. A study from the Inter-American Development Bank found that an earthquake in Mexico led to an 11 percent decline in people’s trust in government.

The U.S. federal government now frequently creates fact-checking and myth-busting websites to counter rumors following disasters. A review of recent research found that false information has often “led to resource misallocation, overwhelmed emergency services, and diverted attention away from genuine needs.”

The terms “misinformation,” generally meaning unintentional deception, and “disinformation,” meaning intentional deception, have been the subject of extensive academic research and policy initiatives. In 2022, the Biden administration appointed a “disinformation governance board” at the Department of Homeland Security. But the terms can be difficult to define in practice: Whether information is considered “disinformation” can depend on the beliefs and attitudes of the person evaluating it.

Even so, the broader problem is clear. Many people believe claims about disasters that are demonstrably false, and they often distrust the people responsible for protecting them. When disaster strikes, that lack of trust can become a matter of life and death.

How unrelated images became a Maui conspiracy theory

The Maui theory grew from a series of unrelated photographs and videos.

On January 16, 2018, the Marathon Petroleum Refinery in Canton, Ohio, conducted a flare in response to a mechanical problem. A local resident photographed the blurry flare, which included an upward-pointing lens flare, and posted the image on Facebook. A local report documented the incident.

In May 2018, SpaceX posted an Instagram photograph of a Falcon 9 rocket launch from Vandenberg Space Force Base in California. The post remains available on Instagram.

In May 2023, a windstorm in Santiago, Chile, caused an electrical transformer to explode. A resident recorded the event, including a lens flare from the bright explosion, and posted the video on TikTok.

In August 2023, a brush fire in Maui, Hawai‘i, spread out of control and destroyed the town of Lahaina. The fires caused more than $5.5 billion in damage and killed more than 100 people.

Soon after the fires began, posts on Instagram, TikTok and X resurfaced the refinery flare, the electrical explosion and the rocket launch. They presented the images as evidence of an energy weapon used to ignite the Maui fire. The posts were viewed and reshared millions of times, often taking on meanings far removed from the original events.

A Hawai‘i-based Instagram user, @for_our_rights_kauai, posted the refinery photograph and suggested that the Maui fires resulted from an “attack.” Several days later, the user acknowledged that the photograph was unrelated to the fires.

“I want to apologize for contributing to spreading that photo that I was told came from Maui on the day of this event. It turns out it was not true,” the user said in a follow-up video. “It is very important for us to vet our information.”

The correction did not travel as far as the original claim. The Instagram account @cultivateelevate, operated by an influential MAHA-adjacent health-food entrepreneur, had already published a video containing a screenshot of @for_our_rights_kauai’s later-deleted post.

“My friend in Hawai‘i,” the host said while pointing to the screenshot, “showed a laser beam coming out of the sky, directly targeting the city.” The video received 432,000 views—more than 40 times as many as the video apologizing for the error.

Other accounts, including the Instagram user @mama_speaks_truth, added to the spread by posting screenshots of screenshots of screenshots. Each successive version moved further from the original refinery photograph, which had been taken five years earlier and 4,500 miles away.

The episode illustrates how disaster misinformation can develop. Images and facts combine with existing fears and suspicions, then become a conspiracy with a new meaning and an independent life.

Distrust of government

Experts have identified several points at which communication can break down before, during and after a disaster. Distrust of government is among the most important.

Federal, state and local agencies and officials forecast weather, issue disaster warnings, coordinate mandatory evacuations and lead emergency responses. If people do not trust those institutions, or if officials cannot communicate effectively, lives are at risk.

The broader trend is unfavorable. Pew Research Center data shows that fewer than 20 percent of Americans say they trust the government to do what is right most of the time, the lowest level recorded.

The Federal Emergency Management Agency is one of the most visible representatives of government during a disaster, making it a frequent target of conspiracy theories and disinformation. Although FEMA is intended to support state and local authorities, disaster-affected communities often circulate claims that federal aid is absent, that the government is restricting freedom of movement or, as in Maui, that officials intend to seize land and disenfranchise victims.

“FEMA has been the source of misinformation since its inception,” said Deanne Criswell, who led the agency during the Biden administration. She cited longstanding allegations that FEMA plans to imprison disaster victims in government-run “FEMA camps.” Those claims have circulated almost since the agency was founded.

“It’s this [idea] of the black arm of government,” Criswell said. “It makes it harder for us to actually get the assistance to the people who really need it, because they don’t know who to trust and who to believe.”

There is no way to know how many disaster victims avoid applying for FEMA assistance because of such rumors and fears. During Hurricane Helene in 2024, misinformation disrupted the agency’s response in North Carolina. A rumor about a man who had come to the state to shoot FEMA disaster-response officials forced Criswell to shut down field operations for several hours during the storm’s immediate aftermath.

Criswell and her deputy chief of staff also became victims of “swatting,” in which anonymous callers make false accusations to police that result in the deployment of SWAT teams. The disruptions added pressure to an aid system that was already strained, likely causing some North Carolina residents to wait longer for assistance and shelter. Reporting by NBC News, The Washington Post and The Los Angeles Times described related threats and disruptions.

Helene struck only weeks before a consequential presidential election, which may have contributed to the spread of political misinformation. Right-wing accounts with names such as End Wokeness accused President Joe Biden of withholding aid from North Carolina, attracting millions of views despite a lack of evidence.

Elon Musk claimed on X that FEMA was preventing him from flying into the state to deliver aid, then later withdrew those claims. His retraction received 1 million views, while his initial accusation received nearly 30 million. The original posts remain available on X and X. The disparity in reach means many people in North Carolina and elsewhere may continue to believe that Biden attempted to restrict aid.

Distrust does not always arise from political hostility. Some people who have experienced storms that failed to match early projections may later dismiss evacuation warnings. Others may fear evacuation because they are undocumented and worry about retaliation. The result can be the same: more people exposed to danger and more people unwilling to seek assistance.

“It could just be a person who says, ‘I’m a conspiracy theorist, I don’t trust the government,’ but there are also people who can’t or won’t trust that government for very different reasons,” Silver said, citing undocumented migrants and communities with traumatic experiences involving law enforcement. “There is some skepticism and doubt, and that is hard to overcome.”

Social media as an accelerator

Given the central role of social media in Americans’ lives, it is not surprising that platforms have become incubators for many modern disaster conspiracies. Rumors about extreme weather are not new. Claims once alleged that wealthy industrialists caused the 1889 Johnstown flood, which killed more than 2,000 people in Pennsylvania. After Hurricane Katrina, some people claimed FEMA placed survivors in concentration camps.

Platforms such as Facebook, Instagram and X have allowed those rumors to circulate globally within minutes. They also make it extremely difficult to track the origin or path of a claim unless researchers at a university or nonprofit laboratory conduct network analysis across millions of posts.

Reliable data on the spread of disaster rumors remains limited, but studies show that the phenomenon is widespread. A study of the 2023 earthquake in Turkey found that about one-quarter of surveyed residents believed at least some version of a conspiracy theory claiming that the United States caused the earthquake with a seismic weapon or bomb. People who believed those theories were more likely to support the incumbent government before that year’s election.

A comprehensive study by the Center for Countering Digital Hate found that most conspiracy posts came from verified accounts. Even seemingly harmless fabrications can cause harm. A doctored image of a shark on a flooded highway deceived media organizations and at least one U.S. senator, according to reporting by BuzzFeed News and The Guardian.

A meta-analysis published last year found that “misleading messages … often outpaced official communications, resulting in reduced trust and exacerbating anxiety, stress, and fear.”

The effects can be physical as well as psychological. A study of a 2018 Japanese earthquake found that communities with low trust in government were more likely to evacuate when evacuation was unnecessary, potentially clogging roads needed by people who genuinely had to flee. After the Maui fires, Criswell worried that fears of a colonial land grab had led some families who needed assistance to reject it.

Although people often blame bots or hostile foreign actors for spreading rumors, some studies suggest that most people who share false claims believe them. Accounts in Russia and China may have fueled outrage about the Maui fires, but many individuals who repeated the misinformation appear to have considered it true.

One prominent source of the Maui conspiracy claims was Cultivate Elevate, an account with more than half a million followers. It is operated by a MAHA-adjacent health-food entrepreneur who posted videos alleging that the fires were deliberately set as part of a government scheme. The account’s owner did not respond to a request for comment. The account’s Instagram page remains available.

The rise of artificial intelligence

The spread of large language models such as Gemini and ChatGPT has introduced another complication. About one-quarter of U.S. adults now use AI chatbots daily, and searching for information is their most common use, according to a Pew Research Center report.

When answering questions, these systems are more likely to rely on news articles and official sources than on social media rumors, which can help counter misinformation. Asked about a “Maui directed energy weapon,” for example, ChatGPT responds that “there is no credible evidence supporting” the theory and offers to explain the origins of images said to show lasers starting the fire.

That does not mean chatbots are reliable in emergencies. A model could retrieve a news article published the previous week saying that a storm evacuation was not necessary, even though conditions had since changed or a new event had developed. It may also be unclear whether a chatbot can accurately summarize someone’s eligibility for FEMA benefits.

Generative AI is advancing faster than most researchers can conduct peer-reviewed studies of its use during disasters. The intersection is nevertheless expanding. Dozens of startups claim to map flood risks or storm damage. The Red Cross has developed a chatbot called Clara to help people navigate disaster situations, while MIT has created a similar system called Crisischat. The Trump administration’s chatbot, America.gov, provides information about obtaining FEMA assistance.

Researchers still know little about how people decide whether to trust these tools or general-purpose chatbots such as Gemini.

“You need to learn if the information’s accurate, but then you also need to know, Do people trust it, regardless of the objective veracity?” said Anita Atwell Seate, a professor of communication and crisis expert at the University of Maryland. “If people trust it, they’re going to act on it whether it’s accurate or not.”

Seate is working with other researchers to develop a longitudinal study examining how people in disaster areas use AI tools. Related research and commentary has been published by RAND, the Red Cross, MIT and America.gov.

Building trust before disasters

It is difficult to define what it would mean to solve misinformation in an information environment as fragmented as that of the 21st-century United States. There is no reliable way to count the number of people who believe a false rumor, and corrections will generally be slower and less viral than the original claims.

The immediate priority is not whether someone in Iowa believes that a laser started the Maui fires. It is whether people in disaster areas can obtain accurate information before, during and after extreme weather. That can determine whether recovery is slow or fast, and in some cases whether people live or die.

Experts generally say the most effective response is to establish strong communication channels before an emergency and have emergency managers work with people who already command trust in their communities.

Near the end of the Biden administration, FEMA began experimenting with that approach. The agency partnered with a Native Hawaiian FEMA employee in Maui who reassured a broader network that the government did not intend to take anyone’s home. Although some people in the United States still believe Biden started the fire with a laser, most people living near the fire eventually sought assistance from FEMA.

FEMA also partnered with the Rosalynn Carter Institute for Caregivers to distribute disaster information through home health aides and caregivers. The goal was to place accurate information with trusted messengers who would have credibility when an emergency arrived.

“From a government perspective, they’re just like, ‘They don’t know us,’” Seate said, describing communication difficulties she has observed at the National Weather Service. “But they don’t have to know us, they know Jim down at the fire station and Jim, you know, rings the alarm and posted on Facebook for us.”

The strategy works best when those relationships are established before a disaster. Even then, it is not guaranteed to succeed. Some people will choose a social media hoax over a trusted meteorologist because the hoax confirms their existing beliefs or for other personal reasons.

Fighting misinformation is not a contest in which one side wins and the other loses. It is closer to trench warfare, with emergency managers and recovery leaders working to expand the share of the information environment occupied by accurate, useful guidance.

“It’s at a level that we’ve never really seen before, because the systems and the silos of information have gotten so siloed and quick,” said Justin Knighten, a risk-communication consultant who led FEMA’s external affairs team under President Biden during several major disasters. “When trust is low, you have a lot of work to not only build up that trust, but to also get ahead of those upcoming threats, so you’re really working at a deficit.”