Hours after the deadly Tejas crash claimed the life of an Indian Air Force pilot at the Dubai Air Show; a video appearing to show Air Chief Marshal AP Singh chastising the government against inducting the indigenous fighter jet into the air force, circulated on X.
It wasn’t real. Singh never made any such remarks. The clip was fake — the air chief marshal’s voice reconstructed or cloned with the help of artificial intelligence (AI).
The fake video was yet another example of a co-ordinated disinformation campaign targeting India since the Pahalgam terror attack in Kashmir and India’s military Operation Sindoor earlier this year.
A new frontier of synthetic propaganda
A cluster of X handles are targeting India in a bid to destabilise the online discourse through a mix of videos manipulated with AI, fake quotes, bogus letters and fabricated news articles.
The Elon Musk-owned platform’s newly introduced ‘About this account’ feature showed their locations as Pakistan.
It is not clear who is behind this campaign, however, it bears signatures of an influence operation from a troll farm - synchronised activity, rapid amplification by seemingly unrelated accounts and a mix of fake personas that mimic Indian X users.
What makes the operation unique is its indiscriminate use of generative AI.
BOOM published more than 30 fact-checks flagging synthetic media, since the Pahalgam terror attack.
The campaign has singled out India’s most visible institutions - its military, its national leadership and the national media.
The influence operation also attempted to exploit moments of national crisis, widen internal faultlines —- religious polarisation and caste prejudice, and fan tensions in sensitive border areas such as Manipur and Ladakh.
New Delhi has responded by pressing X to legally withhold these accounts in India to limit their visibility and by debunking the disinformation through its state-run PIB Fact-Check handle.
Sky wars and the battle of narratives
India and Pakistan have sparred over the number of fighter jets lost during a four-day conflict that broke out in May this year.
Pakistan has asserted it shot down six Indian aircraft. India rejected the claim but acknowledged in an interview, it did lose an unspecified number of fighter jets.
An information battlefield without borders
BOOM identified key X accounts seeding AI-manipulated videos targeting India since the Pahalgam terror attack.
The accounts @InsiderWB, @Baba_Thoka, @Hawkss_eye and @abubakarqassam, featured frequently among debunks published by Indian fact-checkers.
The disinformation posted by these handles were amplified by other suspicious accounts which posted rhetoric pro Pakistan army and the Pakistan Muslim League-Nawaz (PML-N) party.
The Afghan Taliban regime was another target of disinformation and trolling as part of the same campaign.
The tweets were deleted after being called out in many instances. We also observed a few central accounts purging their posts periodically, making it harder to detect their role in the disinformation campaign.
It is difficult to estimate the total number of accounts in the campaign as many handles such as @InsiderWB, @Baba_Thoka were withheld in India or suspended by the platform after being reported.
The disinfo campaign also exhibited signs of inauthentic behaviour. For instance, fake accounts such as @akrittisharma, @TaraSharma02, @Kussikhuelafn tried to camouflage themselves as Indian users. Their profiles had pictures of the tricolour and political slogans. But X’s metadata gave them away showing their location as Pakistan.
Prime targets: Top Indian defence personnel
The top tier of India’s defence establishment have been consistently targeted by the influence operation.
The disinformation narratives include an admission of tactical missteps; territorial concessions to China, fake casualty figures; fabricated statements chastising the government for communalising the army or misusing it for political gains.
Other fake narratives included attacking the secular credentials of the armed forces; undermining the perception of the country’s defence capabilities; and conspiracy theories such as the blast near Delhi’s Red Fort being a ‘false flag’ operation by India.
Videos of defence personnel giving public speeches were overlaid with AI voice clones to make fabricated statements.
Some of the words used in the AI voice clones were Urdu, Pamposh Raina, head of the Deepfakes Analysis Unit (DAU) of the Trusted Information Alliance, in India told BOOM.
The DAU works with a coalition of fact-checkers and experts in synthetic media globally to combat deepfakes.
“...I'd say these were chiefs…chief of the army staff, chief of navy, air force and chief of the defense staff; whenever they made any statement; we would always know that okay this is going to be manipulated and that would happen,” Raina told BOOM.
In November, a doctored video of General Upendra Dwivedi saying that the Indian Army will reduce non-Hindu soldiers by over 50 percent by 2028, went viral.
The deepfake was posted by the handle @Baba_Thoka.
DAU’s Raina said military personnel, who historically avoided the spotlight, have given more interviews and made public remarks post Operation Sindoor, making it easier for bad actors to exploit.
“They seem to be more open to speaking to the media or speaking on public occasions much more than before. So their visibility is higher as a result that content is easily available, as a result manipulation becomes easier,” Raina explained.
She also noted how the insignia and name tags on the military uniforms showed signs of distortion, an anomaly usually seen in AI altered videos.
The death that wasn’t
In a brazen escalation, an account named ‘@Mushk_0’ shared four deepfake videos to claim Sonam Wangchuk died in police custody.
The Ladakhi climate activist is being detained under the stringent National Security Act, at a prison in Rajasthan.
Wangchuk’s wife Gitanjali Angmo rubbished the claim and told BOOM she visited her husband on the day the clips were posted.
The politics of falsehoods
The network also falsely claimed Bihar’s National Democratic Alliance leaders funneled money to Imran Khan’s Pakistan Tehreek-e-Insaaf party.
Deepfakes of independent journalist Ravish Kumar, Firstpost’s managing editor Palki Sharma and NDTV managing editor Shiv Aroor were employed to push the falsehood.
The AI voice-overs stumbled over Bihar Chief Minister Nitish Kumar’s name, pronouncing it as “Neetish Kumar” in more than one deepfake.
Why X is fertile ground for influence operations
India had 22.17 million X users as of October 2025 making it the platform’s fourth largest market, according to Statista.
The disinformation operation comes at a time when Elon Musk and other Big Tech have dismissed or downplayed threats of misinformation and scaled back investments in fact-checking capacity globally.
Moves such as dismantling the platform’s trust and safety team, putting Tweet Deck, a crucial search and monitoring tool and X’s API, out of the reach of researchers and fact-checkers, have made it harder to analyse influence operations.
