As a deadly attack shook the tourist town of Pahalgam in Indian-administered Kashmir last month, triggering military conflict between India and Pakistan, another battle erupted online — a war over truth.
Fake videos using artificial intelligence, recycled war footage and fabricated narratives spread like wildfire across X, WhatsApp, Facebook, and YouTube — fuelling fear, outrage and confusion on both sides of the border.
Among the most viral were digitally fabricated images of strikes.
One viral AI-generated fake claimed to show Rawalpindi Stadium in northern Pakistan reduced to rubble after an attack, racking up millions of views.
Another falsely suggested the Pakistani Prime Minister Shehbaz Sharif had admitted defeat.
"This was electronic warfare," said Raqib Hameed Naik, executive director at the Center for the Study of Organized Hate in Washington DC, who compiled a database of hundreds of misleading posts.
"It was weaponised primarily to fabricate false narratives of military success with fictional visual evidence and to feed hyper-nationalist sentiment, baying for war and more blood," Mr Naik told the ABC.
"The goal was to manipulate public opinion — the war of perception is everything that matters in modern warfare."
The rise of deepfakes and made-up victories
One of the most alarming disinformation trends during the Pahalgam crisis was the rise of sophisticated deepfakes, typically created using AI with superimposed or manipulated video, audio or images.
One AI-generated video appeared to show Pakistan Army spokesperson, General Ahmed Sharif Chaudhary, admitting to the loss of two fighter jets.
"The lip sync was nearly perfect," said Nighat Dad, founder of the Digital Rights Foundation (DRF), a Pakistan-based NGO.
"The only thing that gave it away was the dialect of Urdu and some words that Indians typically mispronounce in Urdu.
"Honestly, it was one of the most convincing deepfakes I've seen," she told the ABC.
The goal was clear: to undermine Pakistani morale and inflate Indian victory.
And it worked, she believed.
Shared across thousands of Indian accounts, the video even made its way onto mainstream news debates before being debunked.
"Misinformation is helping change the narrative, it's helping win wars," Ms Dad said.
Another viral post used video game footage — complete with cinematic music and nationalistic captions — to falsely claim that Indian jets had shot down Pakistani aircraft over Bhuj.
"It was crafted to look like a decisive military win," said Mr Naik.
"But it was just a flight simulator."
Old footage of downed fighter jets was also shared claiming Pakistan victories in the air.
Groups from both countries pointed out that the disinformation campaign was not one-sided.
Pakistan's deputy prime minister and local media also shared a fabricated article purportedly from the UK's The Daily Telegraph praising Pakistan's Air Force as the "king of the skies", despite the newspaper never publishing such a piece.
When disinformation goes mainstream
Both misinformation — misleading content shared without the intent to deceive, often by everyday users — and disinformation — the deliberate spread of false information with the intent to manipulate or cause harm — were prominent during the conflict.
Disinformation didn't just spread from fringe accounts — verified users and even mainstream media amplified unverified content.
In one notable case, a video of a couple dancing on a hillside in Kashmir — originally posted by the couple themselves — was wrongly claimed to be the "final moments" before they were killed in the Pahalgam attack.
Channels like India-based Times Now, The Daily News and Analysis and several others broadcast the video without verification.
The next day, the couple posted on Instagram: "Hey guys, we are alive … we had to delete our original post because it sparked so much hatred."
"Even with the couple debunking the video, it still spread like wildfire," said Sara Imran, a research associate with DRF.
Some in the online community even doubled down when Ms Imran alerted users about the misused video.
"If it psychologically hurts the other side, it doesn't matter if it's fake," one person replied.
Recycled footage fuels narrative battles
In other examples, old footage of a naval drill circulated as an Indian Navy attack on Karachi port while clips from Israeli air strikes in Gaza were passed off as Indian strikes on Pakistan.
These claims sparked panic in Karachi and Peshawar, where residents initially feared imminent attacks.
Once the panic faded, some locals posted videos of themselves drinking tea outside juxtaposed with the sensationalist headlines, mocking the false reports.
"Meme culture became a response," DRF's Ms Dad said.
"Pakistanis countered misinformation with humour, sarcasm, and jokes."
But she also recorded rampant misinformation that fed hate speech in both Pakistan and India and genocidal threats to starve, invade or bomb Pakistan.
Her organisation documented several fake claims of Kashmiri locals sheltering terrorists.
But efforts to counter the falsehoods ran into roadblocks and censorship.
A broken system
On April 28, India blocked 17 Pakistani YouTube channels and several X accounts, including journalists, media outlets like Dawn News and Samaa TV, and even official Pakistani government handles.
The Indian government and X did not reply for a request for comment, while YouTube asked for further details.
"Civil society has long criticised India's blocking regime for its opacity, lack of transparency and absence of due process," Ms Dad said.
"[And criticised] blanket bans based on identity or viewpoint such as targeting accounts for being Pakistani or critical of India."
Despite efforts by community fact checkers and organisations like AFP Factcheck, the scale of misinformation was overwhelming.
BOOM Live, an independent fact-checking initiative, recommended that media outlets implement stricter verification protocols, especially during conflicts.
It also suggested social media platforms enhance their algorithms to detect and flag misleading content more effectively.
"The system of checks and community notes [on X] failed miserably, especially during the conflict," Mr Naik said.
In his research, he found X emerged as the central battleground.
Despite mass circulation, very few posts were flagged or removed across different social media platforms.
The ABC also contacted Meta (the owner of Facebook), but it did not respond in time for publication.
"Of the 437 posts we examined [on X], 179 came from verified accounts. Only 73 had community notes," he said.
Mr Naik tried reaching out to X but received no response.
Even more troubling was the muted response from platform moderation systems, Mr Naik and Ms Dad said.
"It's about civilians knowing the truth and differentiating between what is a lie," Mr Naik said.
"Truth becomes the casualty of war and cross-border disinformation and fact-checking units are needed."
