The Deepfakes Analysis Unit (DAU) analysed a video that apparently shows N.R. Narayana Murthy, co-founder of tech giant Infosys, and Shereen Bhan, managing editor of the CNBC TV18 business news channel, promoting a financial investment project. After running the video through A.I. detection tools and getting our expert partners to weigh in, we were able to conclude that the video was manipulated using A.I.-generated audio.
A Facebook link to the nearly four-minute video in English was sent to the DAU tipline for assessment. The video, embedded in a post, was published on April 2, 2025 from an account with the display name of “Zoie McKenzie” and a display picture of a woman. In their profile details they indicate that the account belongs to an “advertising agency” in California, U.S.A. We don’t have any evidence to suggest whether this suspicious video originated from this account or from another. The video is not available on that link anymore, however, it had garnered at least 70,000 views until a few days ago.
The video appears to be a recording of a public event. It opens with two insets of unequal size placed side-by-side on a blue-coloured backdrop. Ms. Bhan can be seen in the larger inset while Mr. Murthy in the other. The microphone holder in Bhan’s hand carries a logo resembling that of CNBC TV18, while the one in Murthy’s hand has the number “25” emblazoned on it.
Bold, static text graphics in English are visible at the bottom of the video frame, throughout. The first line reads: “My Financial Project For Indians”. The next one claims that, "every Indian citizen who invests 21,000 rupees” in the supposed financial project can get returns worth “1,500,000 rupees in the first month”.
Even as video tracks of Bhan and Murthy play out simultaneously in their respective insets, an accompanying voice can only be heard with Bhan’s visuals. The female voice recorded over her video track rephrases what is conveyed through the text graphics, adding that Indians need to open some sort of an account to make the suggested sum of money. It claims that the project is Murthy’s brainchild and “automatically increases” the funds of users.
The next segment of the video, which continues till the end, shows Murthy in a close-up with some wide shots of him and Bhan on a stage used as cutaways and a few shots of what seems to be an audience. He can be seen seated on a chair, as he holds a microphone and occasionally fiddles with a piece of paper. Judging by the movement of his eyes, it appears that he is not looking into the camera, instead he’s looking at someone or something else.
The backdrop of the shots featuring Bhan and Murthy carry important visual cues about the origin of the video track. In addition to a prominent logo identical to that of CNBC TV18, a logo similar to that of HSBC or The Hong Kong and Shanghai Banking Corporation, a London-based bank is also visible. “Global Leadership Summit” can be seen written in bold. The colors of the Indian national flag, digital representations of high-rise buildings and random stock charts are also visible in the background.
A male voice recorded over Murthy’s video track touts the project as reliable and highly profitable; promising daily earning of “50,000 rupees” and monthly returns to the tune mentioned in the text graphics. It suggests that there are some “participant statistics” that confirm how people’s lives have changed through the supposed financial project, however, no such statistics are shared in the video.
The voice claims that the purported initiative is powered by artificial intelligence and that the "algorithm does all the work for you”. It also declares that within the first month the users’ returns would far exceed their salary, “no matter how high it is and I say this from personal experience.”
A sense of urgency is created by that voice as it asserts that, “registration may close within a few days” because of “high level of interest” and that this “opportunity” cannot be offered to every Indian. It urges viewers to invest “21,000 rupees” to see their money “grow almost daily” and promises that, “early users will gain the greatest benefits”.
Bhan’s lip movements are mostly in alignment with the audio track. However, in some frames her handheld microphone can be seen getting compressed by her chin, in turn distorting the shape of her chin.
In the segment focussed on Murthy, his lips seem aligned to the audio track as long as none of his hands are visibly close to his face. The moment either of his hands appears to occlude his mouth, at any angle, his lip movements get out of sync with the audio. He switches the microphone between his hands and gestures with the other throughout; in some frames when the hand in motion gets closer to the face it seems to blend into the lower part of his face. The microphone appears to merge with his jawline and chin in several frames.
The video quality overall is poor. Bhan’s teeth appear blurred. Murthy’s upper set of teeth have inconsistent shape throughout the video. They go from a brown patch —making it seem that there are no teeth— to a few teeth arranged like a wedge. His lower set of teeth appear as patches of white and brown across different frames.
The voice attributed to Bhan and Murthy, respectively, bears some resemblance to their real voice when compared with their recorded videos available online. However, Murthy’s natural accent does not come through in the audio nor does it carry the pauses characteristic of his delivery. The audio tracks for both sound scripted, lacking natural intonation or pitch.
This is yet another doctored video where an initial investment amount of “21,000 rupees” is being recommended. The DAU has debunked several financial scam videos, such as this, this, and this, promoting dubious investment platforms where the same number was used. Another similarity between this video and previous scam videos is the messaging, which tries to push people to act by conveying that the supposed financial opportunity will be available for a limited window.
We undertook a reverse image search using screenshots from the video being analysed through this report. Both Bhan’s and Murthy’s clips were traced to this video published on the official YouTube channel of CNBC TV18 on Nov. 14, 2024, which does not mention any financial project.
The clothes and backdrop of Bhan and Murthy are identical in the video we analysed and the one we traced. A few clips of Murthy addressing an audience and Bhan interacting with him have been lifted and used to create the manipulated version. There are other people featured in the original video, which too is in English.
In the original video, text graphics at the bottom and toward the right of the frame highlight Murthy's comments and celebrate his presence at an event. In the manipulated video the logos of CNBC TV18 and HSBC are visible only as part of the backdrop. However, in the original video these logos have been superimposed at various points, in addition to a small logo representing a stylised pyramid though we couldn’t discern what or who it represented.
To discern the extent of A.I. manipulation in the video under review, we put it through A.I. detection tools.
The voice tool of Hiya, a company that specialises in artificial intelligence solutions for voice safety, indicated that there is a 31 percent probability of the audio track in the video having been generated or modified using A.I.
Hive AI’s deepfake video detection tool indicated that the video was manipulated using A.I. It pointed out markers of manipulation in a small segment featuring Murthy but did not indicate any manipulation in the clip featuring Bhan. Their audio detection tool did not detect any A.I. manipulation in nearly the entire audio track of the video.
We also ran the audio track from the video through Deepfake-O-Meter, an open platform developed by Media Forensics Lab (MDFL) at UB for detection of A.I.-generated image, video, and audio. The tool provides a selection of classifiers that can be used to analyse media files.
We chose six audio detectors, out of which four gave strong indicators of A.I. manipulation in the audio. AASIST (2021) and RawNet2 (2021) are designed to detect audio impersonations, voice clones, replay attacks, and other forms of audio spoofs. The Linear Frequency Cepstral Coefficient (LFCC) - Light Convolutional Neural Network (LCNN) 2021 model helps distinguish between real and synthetic speech to identify deepfake audio.
RawNet3 (2023) allows for nuanced detection of synthetic audio while RawNet2-Vocoder (2023) is useful in identifying synthesised speech. Whisper (2023) is designed to analyse synthetic human voices.
For a further analysis on the audio track from the video we put it through the A.I. speech classifier of ElevenLabs, a company specialising in voice A.I. research and deployment. The classifier result indicated that it was “very unlikely” that the audio track used in the video was generated using their platform.
We reached out to ElevenLabs for a comment on the analysis. They told us that they were able to confirm that the audio is A.I.-generated. They added that they have taken swift action against the individuals who misused their tools to hold them accountable.
For expert analysis, we escalated the video to our detection partner ConTrailsAI, a Bangalore-based startup with its own A.I. tools for detection of audio and video spoofs. The team ran the video through audio and video detection models, the results that returned indicated A.I. manipulation in the video and audio track.
They stated that the voice’s pacing is unnatural in certain parts and is relatively monotonous compared to real speech. They added that, while there are visual indicators to suggest that the video is A.I.-manipulated as animated lip movements have been synced with the A.I.-generated audio speech, their models fell short on providing the desired visual analysis.
To get another expert to weigh in on the video, we escalated it to our partner Validia, a San-Francisco based deepfake cybersecurity service. They used their proprietary software to check the authenticity of the video and audio being associated with Murthy.
They retrieved a clean sample of the voice purported to be Murthy’s from the video we escalated. Then, they compared that voice with a real voice sample of Murthy using heat-maps. They tested numerous audio metrics and were able to obtain an overall similarity score of 0.93 between a real sample of the target’s voice and the fake provided by us. They explained that a similarity score of 1 would be perfect similarity and anything between 0.8-0.99 would be a likely deepfake voice.
They added that the heat-maps, combined with the score, led them to believe that A.I. technology was definitely used to replicate his voice, although it was not done perfectly.
The team at Validia stated that the video is certainly a better-produced lip-sync deepfake than prior examples with only a few discrepancies. It indicates that this technology is improving and that bad actors are learning to use it more, they added.
Based on our findings and analysis from experts, we can conclude that original footage was manipulated with synthetic audio to fabricate the video. This is yet another attempt to promote a financial scam by falsely linking it to a business leader and a prominent business journalist.
(Written by Debraj Sarkar, and Rahul Adhikari, edited by Pamposh Raina.)
Kindly Note: The manipulated video/audio files that we receive on our tipline are not embedded in our assessment reports because we do not intend to contribute to their virality.
.png)
