Ever since a deepfake video featuring popular Bollywood actress Rashmika Mandanna surfaced online, it has sparked widespread concerns, calling for stronger measures against such deceptive content. Similarly, actress Katrina Kaif found herself a victim of a manipulated video extracted from an action scene in the upcoming movie ‘Tiger 3,’ causing a stir across social platforms. These deepfake videos aimed to tarnish the image of these celebrated actresses.
Following the revelation of Mandanna’s fabricated video, notable figures from Bollywood, including Amitabh Bachchan, rallied in support of the actress, urging authorities to address the issue. Subsequently, the Ministry of Electronics and IT (MeitY) issued advisories to major social media platforms, instructing them to remove misleading AI-generated content, notably deepfakes, within 24 hours.
While Mandanna’s incident caused a storm on social media, it’s crucial to acknowledge that deepfakes are not a recent phenomenon. During the Covid-19 lockdowns, a TikTok user resembling Hollywood star Tom Cruise, known as Miles Fisher, captivated millions with deepfake videos impersonating the actor. These videos, depicting Cruise playing the guitar, golfing, and engaging in TikTok trends, rapidly circulated across social platforms.
Identifying Deepfakes: Recognizing the Signs
Deepfakes leverage deep learning AI to produce falsified images and videos depicting fabricated events, enabling the manipulation of political figures, celebrities, or athletes to mimic actions or statements. Though this technology has positive applications in movies, augmented reality, and other fields, its misuse is prevalent, often contributing to political misinformation, revenge porn, fake celebrity content, and fraudulent schemes.
Creation Process of Deepfakes
The creation of deepfakes involves an algorithm that identifies facial similarities and reconstructs them based on shared features. This process uses decoders trained on different faces. Essentially, a compressed image of person X is fed into a decoder trained on person Y. The decoder then morphs the features of X based on the expressions of Y, resulting in a manipulated video.
Protecting Against Deepfakes and AI Misuse
To combat the dangers posed by deepfakes and AI misuse, it’s crucial to exercise caution and adopt preventive measures. Users should remain vigilant by:
- Verifying Sources: Always cross-verify information and sources before believing or sharing content.
- Raising Awareness: Educating oneself and others about the existence and implications of deepfakes is essential.
- Utilizing Authenticity Tools: Explore tools and software designed to detect potential deepfakes.
- Enhancing AI Regulations: Advocating for stringent regulations and ethical use of AI technologies is vital to curb misuse.
In conclusion, staying informed and exercising vigilance are key to combatting the spread of deepfakes and preserving the authenticity of media in an AI-driven world.