Claim: Videos circulating on social media claim to show Indian Prime Minister Modi laughing at the BRICS Summit with Israeli Prime Minister Netanyahu and Chinese President Xi Jinping.
Fact: This clip is AI-Generated.
On 13 September 2026, Instagram user Pakistanownews posted a clip showing Modi laughing with the other two leaders at the BRICS Summit, as seen below.
Screenshot from the viral Instagram post.
BRICS Summit
The BRICS cooperation is an intergovernmental coalition that was initially founded in 2009 by Brazil, India and China, with more countries joining the coalition through the years, most recently Indonesia in 2025.The core aim of BRICS is to reshape global governance so that Western dominance is reduced, advocating for greater representation in global organisations. Other core principles are to coordinate economic policy, reduce reliance on the US dollar, and create an alternative financial system.
BRICS does this by promoting trade in local currencies than in the US dollar, establishing alternatives to Western institutions such as the New Development Bank (NDB) and Contingent Reserve Arrangement (CRA), and by pushing for reforms in bodies such as the UN Security Council and the International Monetary Fund (IMF).
Fact or Fiction?
We conducted a reverse-image search of the video but found no official clip from the summit that matched its contents.
Analysing the background audio, we noticed an audio and visual inconsistency, as the audio does not match Modi’s actual physical movements or facial expressions.The background noise is inconsistent with the isolated microphone audio and strict protocols of an official diplomatic summit. Moreover, the sound heard in the background is similar to the laugh track on a television show, with an audience chattering and applauding.
AI Detection Tools
Soch Fact Check ran the video through DeepFake-O-Meter for further verification. All but one of the scanners give the video a high score for likelihood of being AI-generated, with the SBI scanner being the only one to give a moderate score.
As seen in the report, seven scanners gave a high score for AI-generation: AVSRDD, LIPINC, FTNC, WAV2LIP-STA, LSDA and Tall. All detected different evidence for AI-generation. The percentage and their functions have been stated below for clarity:
- AVSRDD gave a score of 100% – this is an audiovisual deepfake detection that takes into account the correlation of visual and audio cues by comparing discrepancies between audio and video.
- LIPINC displays a result of 100% – this is used for an advanced AI framework designed to detect lip-syncing deepfake videos by analyzing subtle irregularities in the mouth region. This will explain Modi’s facial movements displaying irregularities when he appears to be laughing, especially in his jaw movements.
- Fully Temporal Convolution Network (FTCN) gave a 99.7% result – it is an advanced video face forgery detection framework.
- Wav2Lip-Sta gave a score of 81% – this is an open-source deep learning framework that synchronises mouth movements in any video to match a new audio track. It detected that the mouth movements have not been synchronised properly with the audio.
- AltFreezing gave a probability of 79.9% – this is a technique identifying abnormal facial motion or frame freezing patterns that exist in AI-generated videos.
- LSDA gave a score of 75% – it is known for deepfake detection based on spatial anomalies in manipulated media that identifies subtle inconsistencies between frames.
- TALL (Thumbnail Layout) gave a score of 71% – this is a high-efficiency video face forgery detection model. This result would have been detected only in the last frames of the video, not the entire clip, which could be the reason for the moderate score.
Results obtained from DeepFake-O-Meter.
We also ran the video through UncovAI, which assessed that the video is AI-generated with a confidence level of 57% but added that the audio is likely to be human with a 96% certainty The mixed result suggests that only a certain section of the video was generated by AI, specifically Modi’s microexpressions and other visual inconsistencies picked up by the detector. Meanwhile, the audio is more likely to be human, and was possibly taken from a live studio audience for a television show.
Results obtained from UncovAI.
Based on the results of these AI detection tools and our own human analysis, we can ascertain that video is AI-generated.
Virality
The video has been shared on Facebook, Instagram and X.
On Instagram, it received 53.3K views and 121 likes.
Conclusion: The video of Modi laughing with other world leaders at the BRICS Summit is AI-generated, based on visual and audio inconsistencies observed, as well as results by AI detection tools.
To appeal against our fact-check, please send an email to appeals@sochfactcheck.com