The flood did not announce itself with a single dramatic roar. It arrived as rain, persistent and gray, drumming on tin roofs and muddy paths, then turning from nuisance to menace as rivers began to swell. On an ordinary Wednesday, the rivers of Nepal’s northern valleys are not villains. They are sources of life, carrying meltwater from the Himalayas to fields and villages, sustaining farmers, pilgrims, and the rhythm of everyday existence. But the Wednesday that arrived in late August was not ordinary. Torrential rain turned the Bhotekoshi and Trishuli rivers into swollen, furious currents that smashed through settlements, tore apart bridges, and severed the roads that connect Nepal to its northern neighbor. In the space of hours, families lost their homes, their livelihoods, and in too many cases, the people they loved. As rescue workers struggled to reach isolated communities, another disaster was unfolding in the digital world. On platforms like Facebook, Instagram, and TikTok, disturbing videos and images began circulating: alleged footage of the floods, supposedly real-time scenes from the disaster zone. But much of it was not real. Some clips were old recordings from earlier floods, stripped of context and re-labelled as if they were happening again. Others were created by artificial intelligence, engineered to look like horrific scenes from the affected areas. Still worse, scammers began circulating fake donation QR codes, exploiting public empathy to steal money intended for victims. The government of Nepal responded with a firm demand: Meta and TikTok must remove this harmful material immediately. This was not an overreaction or a bureaucratic gesture. It was a recognition that in an emergency, disinformation is not merely annoying; it is a weapon that can cause panic, block aid, and deepen the trauma of families already drowning in grief. When people cannot trust what they see, they cannot act wisely. When they cannot act wisely, more lives are endangered. The digital storm and the natural one had become one single emergency, and Nepal was determined to confront both.
These are not small bureaucratic details. The person responsible for coordinating Nepal’s official response to this digital chaos is Bijay Kumar Roy, the coordinator of the Government Task Force for Regulation and Management of Social Media. He has been candid about the urgency. In meetings with representatives from Meta and TikTok, Roy and his team did not politely request a review of flagged content; they demanded immediate removal. According to a report by The Kathmandu Post, the government also asked the social media giants to optimize their algorithms so that they could detect and block engineered visuals and false information from the very beginning, instead of waiting for users to stumble upon them. The phrase “from the very beginning” matters. It means the platforms should be using their vast technical and financial resources to prevent harm before it spreads, not after it has already hurt someone. There is a reason for this insistence. Right now, when someone reports a harmful post, Meta and TikTok typically take somewhere between 24 and 48 hours to respond and follow up. In normal times, two days of delay might be tolerable. But in the middle of a flood, with families trapped on rooftops and rescue teams trying to make life-and-death decisions, 48 hours is an eternity. A false video shared during that window can reach hundreds of thousands of people, convince them that a particular road is safe when it has vanished, or persuade them that a relative has survived when there is no evidence. Worse, a single piece of horrifying but fake footage can trigger a wave of panic that overwhelms local helplines and makes it harder for responders to identify the people who genuinely need help. The government’s demand reflects a deeper frustration: platforms should not be passive hosting services in times of disaster. They are some of the most powerful communication tools in history, and with that power comes a responsibility to act as quickly as the emergency itself. In a country where officials are already working under extreme pressure, having to chase viral lies on top of rescue operations is an impossible burden. That is why Roy’s team is not leaving the matter to chance. They are pressing for a new standard of accountability, one that treats misinformation with the same seriousness as the flood itself.
It is worth looking closely at the toxic material that prompted such a strong response, because it reveals how modern disinformation operates. The first category is old video recordings. In the days after the floods, social media users began sharing footage of previous storms and river overflows, presenting them as if they were new images of the disaster that was unfolding before their eyes. Some of these clips were years old. They had been recycled, stripped of dates and locations, and dressed in captions like “Breaking: Bhotekoshi River destroys villages.” The result was a distorted picture of the scale and location of the damage, making people in unaffected areas believe they were in danger, while drawing attention away from communities where the danger was real. The second category is even more sinister: AI-generated visuals. Thanks to powerful tools that can fabricate photorealistic images and videos, anonymous creators produced footage that looked like authentic disaster scenes but had no connection to reality. A computer algorithm could invent collapsed bridges, drowning families, or raging floodwaters in places where the river was still calm. This is not harmless deception. It feeds on human sympathy and fear, and it forces emergency managers to waste precious time verifying the visual rubbish that should never have been published in the first place. The third category is the most despicable: donation fraud. While generous people around the world were looking for ways to help Nepal, scammers circulated QR codes linked to fake charity accounts. They asked for small contributions, the kind that individuals can afford, and then pocketed the money. Every dollar stolen in this way was a dollar that never reached a shelter, a meal, or a medicine pack for a flood victim. The people who created these schemes understood exactly what they were doing. They were exploiting the most beautiful human instinct—the desire to help a stranger in need—and turning it into profit. That is why the government did not merely ask for content moderation. It asked for algorithmic prevention, a system capable of stopping these lies before they ever entered the public square.
To understand why the government had to act so forcefully, one must understand the scale of the disaster that triggered this digital crisis. Since Wednesday, August 26, floods and landslides have hammered northern and central Nepal. The Bhotekoshi and Trishuli rivers, usually fed by steady mountain snowmelt, swelled beyond memory after days of relentless rain. Their overflow inundated residents’ settlements, reducing rows of houses to piles of splintered wood and rubble. Bridges that once connected people to schools, clinics, and markets collapsed into churning brown water. The main road network, the thin asphalt lifeline that links Nepal’s capital to the remote Himalayan regions and onward to China, was cut in multiple places. For thousands of people, this meant more than inconvenience; it meant isolation, with no way to evacuate and no way for aid to reach them. The disaster did not stop at the border. In China’s southwestern region, particularly in Gyirong County in the Tibet Autonomous Region, mud and landslides struck the Gyirong Port, a crucial crossing point between the two countries. Communication networks were knocked out, road access was severed, and electricity supply failed. In the affected mountain valleys, the modern world suddenly vanished: no phone signal, no internet, no power, only the sound of unstable slopes slipping under the weight of water. Families separated by the border on two sides of the Himalayas were equally terrified, equally uncertain about the fate of their loved ones. The physical destruction is staggering, but the emotional toll may be even worse. Imagine waiting for news of someone you love, staring at your phone, refreshing social media only to find a random video that may or may not show their village. In that state of desperation, a single AI-generated image of destruction can feel like evidence. It can make you choke with grief, even when no one has confirmed a single death. This is the cruel atmosphere that the government is trying to clear.
Amid all this darkness, there are stories of ordinary courage. According to official reports, hundreds of local residents, construction workers, and tourists in the affected areas have been evacuated to safer locations. Some of these evacuations happened before the worst of the water arrived, thanks to neighbors who pounded on doors and shouted at families to leave. Others happened in the brutal hours after the floods, with rescue teams pulling people from rooftops and clinging trees. On both sides of the Nepal-China border, joint search-and-rescue teams are continuing their operations along the rivers, scanning the mud-choked banks for signs of life and for the bodies of those who were not so lucky. These teams are made up of human beings who have chosen to risk their own safety for strangers. They wade through chest-deep water, listen for cries beneath debris, and use every tool available, sometimes even their bare hands, to dig through mud and rock. The work is exhausting and heartbreaking, and it is made far more difficult when disinformation floods the airwaves. Every false report of a missing person sends rescuers chasing a phantom. Every fake video of a collapsed bridge can cause a driver to attempt a dangerous detour. Every hoax donation campaign drains resources from legitimate relief efforts. This is why the Nepalese government’s request to Meta and TikTok is so important: not because social media is the enemy, but because in a rescue operation, accuracy is as essential as water and food. When people know the truth, they can make wise decisions. They can avoid dangerous routes, send donations to vetted organizations, and trust the official warnings that tell them where to go and where to stay away. When lies fill the same channels, the entire system of survival begins to break down. The humanitarian response is not only about boats and helicopters; it is about information, hope, and the ability of communities to rally around something real.
Perhaps the greatest lesson from this disaster is that caring for people today requires caring for the platforms they rely on. The government of Nepal has stepped forward with a demand that should resonate far beyond its borders. A social media company that can serve personalized ads in milliseconds can also learn to flag fake donation codes and manipulated disaster videos in minutes. The excuse that moderation takes 24 to 48 hours is not a technical limit; it is a choice. This moment is a test for Meta and TikTok, and they have not yet passed. In the coming days, their actions will show whether their algorithms can be forces for safety, not just engines for engagement. The Nepalese officials have made one thing clear: in an emergency, proactive action is not optional. It is the only ethical response. But this is not only about corporate policy. It is about the rest of us, too. When we see a shocking video of a distant flood, we should pause before sharing. When we encounter a QR code asking for donations, we should verify the charity before scanning. When we feel panic rising from our feeds, we should remember that some stories are manufactured to provoke that exact feeling. Our own digital behavior can help slow the storm of lies. For the people of Nepal, the road ahead is long and painful. They will need to bury their dead, rebuild their homes, and find a way to trust the terrain that has betrayed them. They will not ask for perfection from the outside world, only for solidarity that is real, honest, and informed. And as the water slowly recedes, the mud on every street will dry, but the memory of the lies that circulated during their darkest hours will remain. The hope is that this disaster will force a change: not only in algorithms, but in the way we understand our duty to each other in a world where information can save lives or end them. The rivers may rage again. The next emergency is already waiting somewhere beyond the horizon. The question is whether we will be ready to tell the truth about it, and to protect those who need it most.