Results Summary

• Audio: Sudha Murthy’s Audio appears to be AI generated using RVC cloning techniques.
• Video: Frames are low resolution & predictions are not reliable
• Slowed-down snapshots however indicate AI manipulation

A. Detailed Audio results

• We clip the file and extract the final 10 seconds of the video file
• Both Audio Spoof detection models predict the Audio file to be AI generated with high confidence.
• The audio sounds very similar to original voice, which gives strong indication that Retrieval Voice Conversion (RVC) based cloning has been done to emulate their voice.

Audio Analysis: Sudha Murthy’s Audio appears to be AI generated using RVC cloning techniques

Full PDF report

Full PDF report

B. Detailed Video Results

• Frame deepfake detection model is confused because of the low resolution frames, it is confused between real & fake