A video circulating on social media falsely claims to show an official Australian police briefing announcing the arrest of four Indian Muslims due to their involvement in the recent Bondi Beach terrorist attack.
BOOM found the claim to be false. Our analysis using deepfake detection models indicated a high likelihood of synthetic manipulation. Furthermore, the original footage confirms the police official made no mention of arresting Indian Muslims during her speech.
The disinformation campaign follows the December 14, 2025, shooting at Bondi Beach. While official reports identified two suspects—an Indian immigrant, Sajid Akram, and his Australian-born son, Naveed Akram—authorities have clarified that the radicalisation likely occurred locally in Sydney.
The Claim
The video was shared on X (formerly Twitter) by a user named Perlina Trump (archived here), claiming that Australian agencies had uncovered a network of four Indian Muslims involved in the recent Bondi Beach terrorist attack. The post also alleged that the Indian government issued a statement confirming they are under investigation until sentencing and that "IDF Mossad" had investigated the connection.
What We Found
BOOM found that the viral video is a deepfake created using AI-generated audio and altered visuals.
Original video does not contain viral statement: BOOM traced the original footage to a press briefing published by The Guardian on December 18, 2025. In the authentic video, the police official provides a list of 59 charges against the gunman. We reviewed the entire speech and confirmed there is no mention of the arrest of "four Indian Muslims" or any involvement of the Indian government or Mossad.
Shared by a "Parody AI" account: The account that posted the video, @perlinaino, includes the description "Parody AI Account" in its bio.
Unusual mechanical movements observed: Upon close analysis, we observed unusual mechanical movements on the speaker's face while speaking, which indicated a high likelihood of the video being AI-generated. The lip movements do not naturally align with the audio track during the specific claims about the arrests, a common characteristic of deepfake manipulation.
Deepfake detection tools flag AI manipulation: We and ran the viral video through several models on Deepfake-o-meter, a forensic tool developed by the University at Buffalo. A majority of the models returned a high confidence score for AI generation.
Indian police rules out radicalisation in India: Credible news reports from The Guardian and Times of India confirm the suspects are Sajid Akram and his son Naveed Akram. Telangana Police in India investigated Sajid's background and explicitly ruled out radicalisation in India, stating it occurred entirely abroad.
