Two videos are circulating on social media with the claim that they show Indian journalists Palki Sharma and Shiv Aroor reporting on Bihar NDA leaders being under scrutiny for funnelling money to Pakistan’s PTI.
BOOM found that both clips are doctored using AI, with no such reports appearing on the official platforms of their respective news outlets.
The fabricated videos appear to be part of a wider pattern of AI-manipulated media featuring Indian news personalities delivering politically sensitive statements. A similar video with media personality Ravish Kumar has been debunked by BOOM.
The Claim
Multiple social media posts shared two videos claiming that Firstpost’s managing editor Palki Sharma and NDTV managing editor Shiv Aroor reported that four senior Bihar NDA leaders were being investigated for financing Pakistan’s PTI.
A Facebook post shared both the videos, including the AI-manipulated Ravish Kumar clip BOOM has previously debunked.
What We Found
BOOM found that both viral videos are digitally manipulated and contain AI-generated audio and altered visuals.
- No such report by either anchors or their news outlets: A review of Firstpost and NDTV’s official websites, YouTube channels, and the social media accounts of Palki Sharma and Shiv Aroor revealed no report mentioning any investigation into Bihar NDA leaders funding Pakistan’s PTI.
- Discrepancies in eye movements and speech: Both clips showed unnatural facial expressions, and repetitive eye movements and speech patterns that do not align with natural human motion, indicating the use of AI-driven visual manipulation.
- Deepfake-o-meter flags high likelihood of AI manipulation: BOOM ran the clips through the Deepfake-o-meter tool developed by the University at Buffalo, and a majority of its audio and video detection models flagged both the visuals and the soundtracks as having a high likelihood of being AI-generated.
- DAU analysis confirms AI manipulation: BOOM consulted with our partners at the Deepfakes Analysis Unit (DAU), who ran the clips through multiple audio and video detection tools. DAU flagged repeated upper-body loops, unnatural blinking, and facial distortion. Furthermore, tools including Hive, Hiya, and Aurigin AI indicated a strong likelihood of synthetic manipulation across both visuals and sound.