Hindi news journalists with millions of subscribers on YouTube are grappling with a unique challenge.
Rather than fretting over news stories or algorithm-driven engagement, they now have to worry about a growing crop of fake YouTube channels cloning their voices and propagating disinformation in their names.
An investigation by Decode found over 40 YouTube channels impersonating the voices of popular Hindi news journalists such as Ravish Kumar, Abhisar Sharma, Sudhir Chaudhary, Rajat Sharma and Shubhankar Mishra among others, using AI voice overs.
The admins earn through ad revenue and in some cases charge subscribers for ‘members-only’ exclusive content.
Some of the fake channels have millions of followers and an equally impressive number of views on their videos.
YouTube suspended 21 fake channels after we flagged it to the Alphabet-owned platform. We found another 22 channels over the course of our investigation.
"Every day a new channel is created in my name using my photo and voice. Although people who watched me were confused earlier, now they have started knowing which is my real channel," independent journalist Ravish Kumar told Decode.
The former NDTV Hindi anchor’s real YouTube channel currently has over 14 million subscribers.
Fake Channels Eat Into Follower Base; Warp Perception Of Journalists
Abhisar Sharma, who has over 9 million subscribers on YouTube, explained how a wave of fake AI-based content impersonating him eats into his following, misleads people on important issues and distorts the audience’s perception of him.
“People come to my real YouTube channel and start abusing me. They even believe that I am running that fake channel,” Sharma told Decode.
“The channels posting AI-based content, make a lot of absurd and false claims, which create a lot of problems for us,” he added.
The AI Slop-Disinformation Intersection
The channels use actual videos of journalists and overlay them with AI voice-overs..
A channel called JSR News, impersonated Rajat Sharma, chairman and editor-in-chief of India TV. It routinely shares disinformation while attributing it to him.
When the Taliban-led Afghan foreign minister visited India in October, JSR News and another channel called Bharat Decode TV claimed Afghanistan handed over its Bagram airbase to India and that the latter had hoisted the tricolour there.
Similarly, a channel named Mdy News, with over six lakh followers, posts fake AI-voice-clone based clips of Ravish Kumar taking a pro-Palestine stance.
Another YouTube channel called Sone Chandi Ke Bhav Live (Gold and Silver Prices Live) features an AI voice of Sudhir Chaudhary and claims to provide daily prices of gold, significantly different from market rates.
The channel has 2.15 million subscribers and 2400 videos.
The Delhi High Court recently passed an order protecting the personality rights of Sudhir Chaudhary and ordered the removal of his deepfake videos from Facebook, Instagram and YouTube. Despite this, several deepfake videos of Chaudhary remain available on YouTube.
The Challenge Of Regulating Deepfakes
The intersection of AI-generated media with scams, disinformation, non-consensual imagery, trolling on one hand, and its use for legitimate purposes on the other hand, make the landscape of synthetic content tricky to regulate.
India's Ministry of Electronics and Information Technology (MeitY) recently released draft amendments to the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021, to tackle the growing challenges of AI-generated misinformation, deepfakes, and synthetic media
The proposals include labelling, embedding identifiers, and user verification in relation to synthetic content.
On its part, YouTube launched its likeness detection technology last month that allows creators in the YouTube Partner Program, to request the removal of AI content that misuses their likeness.
Some of the journalists we spoke to said the feature has not been rolled out to them yet.
India TV editor-in-chief Rajat Sharma has filed a personality rights suit and is seeking the removal of deepfakes impersonating him.
Not just journalists, but several Bollywood actors and singers have moved court this year to ring-fence their likeness against misuse using AI.
“A staggered approach seeking to balance innovation and user’s safety with regulations aiming to handle different functional problems is a fair approach,” Arya Tripathi, partner at law firm Cyril Amarchand Mangaldas, said.
“...attribution of liability will be dependent on the relevant facts at hand,” Tripathi said.
Removing Fake Channels - A Game Of Whack-a-Mole
Abhisar Sharma said despite YouTube taking action against fake channels impersonating him, new ones keep cropping up.
"I even complained to YouTube about the AI channels. YouTube takes action, but after some time, many other channels are created again,” he said.
“I am constantly trying to get them removed somehow, but there are so many channels that I don't understand who is running them,” an exasperated Sharma said.
