Combating deepfakes

ADOCTOREd video of actor Rashmika Mandanna has put the spotlight on the dangers of deepfakes and gaps in the regulatory mechanism. Deepfakes are fake media generated by Artificial Intelligence (AI), which is used to alter images and videos to make it appear that some persons are saying or doing something without them actually doing it. Its potential for spreading misinformation and violating privacy is alarming. The morphing tools can be used to commit crimes, harm reputations, influence polls and undermine trust in democratic institutions. A report says adult content targeting mostly women accounts for 98% of all deepfake videos online. India is listed as the sixth most vulnerable country.

The actor’s morphed viral footage has triggered calls for legal action against the culprits. The Centre has been prompt in instructing social media platforms to remove such content within 24 hours of receiving a complaint. Those impacted by the menace have been advised to get FIRs registered and inform the platforms. The IT Act, 2000, provides for a jail term up to three years and a fine of up to Rs 1 lakh for anyone who uses a communication device or a computer resource to cheat by ‘personating’. Taking the culprits to task is essential to gain public trust.

It is predicted that five lakh video and audio deepfakes will be shared on social media sites across the world this year. Several countries have been proactive in addressing the problem. The UK’s Online Safety Act criminalises sharing deepfake porn. China has banned the production of deepfakes without user consent and mandated altered content to be marked as modified. South Korea has made it illegal to distribute deepfakes that could ‘cause harm’ to public interest. India, too, must scrutinise the existing provisions to mitigate the risks.