*Updated on May 15, 2024: The Deepfakes Analysis Unit has revised its set of definitions for labelling synthetic media, based on that we have also revised the label that we had initially used to categorise the video assessed in the report below. We have relabelled it as a cheapfake, the earlier label was deepfake.
The Deepfakes Analysis Unit (DAU) analysed a short video that appears to have been put together using cheap technology to create the likeness of veteran Bollywood actor Amitabh Bachchan offering a strange remedy. We put the video through A.I. detection tools, which confirmed that Mr. Bachchan’s face and voice in the video were a creation of A.I.
The 11-second video clip in Hindi was sent to the DAU tipline for assessment. In this clip a passport-photo-size video still of Bachchan’s appears to have been used in the foreground and moving visuals of cardamoms in a white dish can be seen in the background.
The video still of him seems very robotic with the eyes blinking strangely, the neck barely moving but the lips move constantly and the audio that can be heard seems to pass off some strange recipe involving condiments as a cure. The voice sounds like the actor's but has a robotic tone to it.
The Instagram handle, which posted this clip has posted several other similar clips of the actor patched together using inexpensive tools to peddle egregious recipes as treatment, the language in those videos is also crude.
We put the video through A.I. detection tools to determine if the video had elements of A.I. in it.
HIVE AI’s deepfake detection tool indicated A.I. manipulation in a good chunk of the video.
We also used TrueMedia’s deepfake detector, which overall categorised the video as “highly suspicious”, calling attention to a high probability of A.I. use in the production of this video. As the tool further broke down the analysis, it gave a 100 percent confidence score to face manipulation, deepfake face detection, and A.I.-voice detection — all three indicators that the video is a creation of A.I.
The voice detection tool of Loccus.ai, a company that specialises in artificial intelligence solutions for voice safety, gave results that indicated that most of the speech in the video is synthetic.
We also ran the video through Itisaar, a deepfake detection service, which is a collaboration between IIT Jodhpur and DigitID, a tech startup which has partnered with the DAU. Their analysis gave a .998 confidence score to the video being fake, which according to their metric is a very high score for A.I.-generated content.
We checked Bachchan’s social media handles but found no comments regarding this video that we analysed or the other fake videos of his posted from the Instagram handle that carried this video.
On the basis of our findings, we can conclude that the actor did not say what is being attributed to him through this video and that the visual element of his seen in the video has been produced using A.I.
(Written by Debraj Sarkar with inputs from Areeba Falak, edited by Pamposh Raina.)
Kindly Note: The manipulated audio/video 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)
