PM Narendra Modi gives AI warning, calls deepfakes "problematic"; know how to spot them and to stay safe

On Friday, November 17, Prime Minister Narendra Modi highlighted the galloping problem of deepfakes in India. For the unaware, deepfake is an artificial intelligence (AI) technology where media, such as photos, videos, and audio are hyper-realistically manipulated to make it seem extremely real even though it is fake. The latest case was of actress Rashmika Mandanna, who became a victim of the latest such horrific attack. PM Modi was addressing journalists at the Diwali Milan program at the BJP headquarters in New Delhi.

During the address, he flagged the misuse of AI for creating 'deepfakes', and said the media must educate people about this crisis. The issue of deepfakes has even caused some celebs to take action in court. Earlier this year, actor Anil Kapoor successfully fought a lawsuit against unauthorized deepfakes of himself, and recently, a disturbing deepfake video of actor Rashmika Mandanna came to the surface where her face was added to another woman’s body.

The problem of deepfakes

In many ways, the Rashmika Mandanna deepfake row started the conversation in India around this problem which has the potential to blow up anytime. In this incident, a small six-second clip of the actor was shared online where Mandanna can be seen entering a lift. It quickly became viral. But later, it was revealed that the video was of an Instagram influencer Zara Patel, and Mandanna’s face was added through AI.

In his reaction, Union Minister Rajeev Chandrasekhar said, “Govt is committed to ensuring Safety and Trust of all DigitalNagriks using Internet”. Calling deepfakes the latest and extremely dangerous and damaging form of misinformation, he explained that it “needs to be dealt with by platforms”.

Patel, the woman whose video was deepfaked by bad actors, stated on her Instagram account and said, “I'm deeply disturbed and upset by what is happening. I worry about the future of women and girls who now have to fear even more about putting themselves on social media. Please take a step back and fact-check what you see on the internet. Not everything on the internet is real”.

How to spot deepfakes

The Massachusetts Institute of Technology (MIT), which has its own dedicated AI and ML research department, has published some helpful tips that people can use to differentiate between deepfakes and real videos. A few of them are listed below.

1. Pay attention to the face. High-end DeepFake manipulations are almost always facial transformations.

2. Pay attention to blinking. Does the person blink enough or too much?

3. Pay attention to the lip movements. Some deepfakes are based on lip-syncing. Do the lip movements look natural?

The risk of being deepfaked is low for most individuals due to the extensive training data required to create such sophisticated manipulations. Without a vast collection of personal photos and videos readily available online, it becomes challenging for AI models to produce flawless deepfakes, particularly when lateral facial views are involved.