Rashmika Mandanna, Katrina Kaif, now Kajol: Why are we seeing sudden uptick in deepfake videos

A video purportedly depicting Bollywood actress Kajol changing into an outfit that went viral on social media earlier this week, has been exposed as a deepfake. The misleading video, widely circulated on major social media platforms such as Facebook, X (formerly Twitter), and YouTube, was not originally of the actress but that of a social media influencer.

This revelation comes amid a rising wave of concerns regarding deepfakes, triggered by a previously viral video falsely depicting popular actress Rashmika Mandanna who had something similar happen to her, as well as Katrina Kaif, whose deepfake was made using a scene from here recent film.

A fact-checking platform has confirmed that the deepfake video featuring Kajol originally featured a social media influencer. Kajol’s face digitally manipulated into the footage. For a split second in the morphed deepfake video, the face of the actual woman becomes visible.

The report indicates that the initial video was originally uploaded on TikTok on June 5 as part of the popular “Get Ready With Me” (GRWM) trend, but the original creator remains unidentified.

Deepfakes, which involve not only manipulated videos, but images and audio as well, are becoming increasingly prevalent, mainly in the form of explicit content and political messaging, which is then used for misinformation.

This poses a significant challenge to societal norms. Following the uproar over the deepfake involving Rashmika Mandanna, the Indian government had issued an advisory to major social media platforms, urging swift action against deepfakes and the removal of content spreading misinformation.

Platforms are expected to eliminate such content within 36 hours of a filed complaint. However, existing regulations regarding AI and deepfakes may prove insufficient.

Aaron Bugal, Field CTO at Sophos, highlights the role of advanced AI in the creation of deepfake content, often derived from publicly available social media profiles. While restricting profiles to private settings may reduce exposure, reposting or misuse by known contacts remains a concern. Bugal emphasizes the significance of the advisory from the Indian Ministry of Electronics and Information Technology (MeitY), cautioning social media companies about potential penalties, including the loss of safe harbor rights, for failure to remove deepfake content promptly.

As a protective measure, Bugal suggests the use of digitally signed videos to verify trustworthy content, likening it to certificates validating website security and email communications. With the continuous evolution of technology and improvements in deepfake quality, distinguishing between authentic and manipulated content may become increasingly challenging, necessitating a reliable validation process.

Celebrities, who thrive on a public persona to promote their brand, face particular challenges in mitigating deepfake risks. Bugal recommends proactive monitoring using tools like Google Alerts and BING News Alerts to instantly notify individuals when their name or brand is mentioned online. This allows for prompt investigations and responses to false representations, either through takedown actions or public corrections.