India and neighboring countries are waking up to the threat. Until recently, there were no specific laws against deepfakes, relying on vague IT Act Section 66E (Violation of privacy) or 67 (Publishing obscene material).
Detecting whether a video is a "real" recording or an AI-generated deepfake requires a keen eye for subtle inconsistencies in human anatomy and environmental lighting. As deep learning models improve, these "desifakes" (a term often referring to South Asian-focused deepfakes) become harder to spot, but they almost always leave behind digital artifacts. 1. Inspect Facial Features and Movement
South Asian cinema often features high-contrast lighting, heavy jewelry, and unique facial features (bindi, nose rings, etc.). Western deepfake models often fail here, but specialized DesiFake models have been trained on Telugu, Tamil, and Hindi cinema datasets, making the fakes regionally accurate.
Small-town India has gone digital. Kiranawalas (corner stores) now use WhatsApp to take delivery orders. must capture this "Jugaad" innovation—the art of finding a low-cost, clever solution to a complex problem.