Indian politicos wake up to deepfakes. Can they fix themselves first?

India has suddenly woken up to the menace of deepfakes. Over the past few days, sleazy deepfake videos of actresses Rashmika Mandanna, Katrina Kaif, and Kajol went viral on social media. Instagram is teeming with Reels where Prime Minister Narendra Modi can be heard crooning some gut-wrenching emo tracks.

I'll start with the basics here. Here's how MIT Sloan defines it: "A deepfake refers to a specific kind of synthetic media where a person in an image or video is swapped with another person's likeness."

You can swap a person's dynamic face lookalike on another person's body or synthesize a fake audio based on another person's audio signature. It's scary, and thanks to the recent advances in deep learning and generative AI, it's hard to separate reality from machine fiction.

Remember that viral image of the Pope in a drip that would put a rapper to shame? That's just a start, but it's not exactly a new phenomenon. A guy moving and talking eerily like Tom Cruise is fun to watch. It's scary good, actually. But when you fake someone like US President Obama or at-war Ukraine leader Zelensky? All hell breaks loose.

Even Modi ji's intellect fell to deepfakes!

In fact, our dear own Modi ji has fallen victim. "I saw a video in which I was doing garba," he said at a rally earlier this month. Well, it was not a deepfake as our supreme leader claimed. Instead, it was a businessman named Vikas Mahante. "I would like to address you all that the person in the viral video is me," he later clarified.

That's our leader falling prey. A savant with a degree in entire political science, French proficiency, insights into radar warfare, sewer aka naala gas innovation, and of course, deepfakes. With such a high-caliber deception and even Bollywood getting tangled in the web, our ministers finally kicked into action.

"We plan to complete drafting the regulations within the next few weeks," information technology minister Ashwini Vaishnaw said at a recent meet. The rules will focus on detecting deepfakes (gasp), preventing them from going viral, strengthening reporting mechanisms, and spreading awareness about the technology.

Union Electronics and Information Technology Minister Rajeev Chandrasekhar soon offered more details. "MeitY and the Government of India will nominate a Rule Seven officer and will take a 100% compliance expectation from all the platforms," he said. This Rule Seven officer will be tasked with upholding all the aforementioned codes.

The execution gulf

It all sounds good. But if you've ever tried to lodge a complaint before the cybercell, or heck, even had to visit a police "thaana," you know the drill.  Over 25,000 cyber fraud cases filed were until August this year, a jaw-dropping 212% increase.

All that amounted to a delightful financial loss exceeding Rs 200 crore this year. But here's the worst part. Only 2-8% of the cheated money has been recovered, because clearly, we're nailing this whole cybercrime thing. And here's the scariest part.

Unlike payment scams, you don't need to learn online tools or channel your inner Shah Rukh Khan to dupe someone on a call. Deepfakes tools are easily available – for free – on the web and app stores.

The threat is so real that Microsoft pulled VALL-E from public release due to extremely high risks. That's because this tool only needs a three-second audio sample of someone's voice to start speaking anything just like them.

On the other hand, we have Indian laws covering online crimes. Simply put, they put the onus on the victim. A victim of deepfake crime has to reach the police station to get a report filed, and then reach out to the social media platforms to get the objectionable content pulled.

Both are titanic challenges with a ton of variables. Will they find your evidence enough to file a chargesheet? Is your complaint regarding someone impersonating you in a fake video worth a serious crime? But even if the culprit is caught, those deepfake photos and videos have already been widely circulated.

Will the police reach out on your behalf and tell Instagram to remove those videos? Will it ask WhatsApp to prevent the sharing via its communications platform? For an average person without resources and legal appetite, that's like pushing a mountain.    

In the case of Indu Jain v. Forbes Incorporated, the Delhi High Court made the rather astute observation that celebrities enjoy the luxury of effectively addressing any damage to their reputation within the marketplace of ideas, whereas individuals with limited resources may find themselves at a disadvantage when attempting to do so.

But there's more here. India's own penal code has no special provision for deepfakes. In fact, under the existing laws, deepfakes would be deemed harmful if they lead to monetary loss. Such is the state of affairs that in the contentious IT Rules (2021), which have been in the formulation phase for years, we still don't have any concrete end-to-end guidelines to deal with deepfakes.

The Ministry of Electronics & Information Technology has asked social media platforms to remove deepfakes within 36 hours of reporting them. But once again, that only happens if you register an FIR and the police take a proactive approach to address your plight. Back to square one!

The social media shit-show

But let's say you take matters into your own hands and try to rely on a social media platform's own reporting system. Well, that chain of actions has historically been sloppy and unreliable, unless you're a high-profile celebrity that puts Meta's own image at serious risk.

But even they're not safe. According to an investigation by the Center for Countering Digital Hate, Instagram “systemically fails” at safeguarding high-profile women from the “epidemic” of misogynistic abuse. Welsh model Jess Davies recently highlighted how Instagram has made it even more difficult to report fakes.

Facebook already has its hands full of unique forms of sexual abuse content problems. But India is a special case in this situation. The ominous kind of special, to be specific. This report from The Washington Post shows how Meta has given the cold shoulder to India when it comes to addressing problems like hate speech and violent content.

A Vice report exposes how Facebook failed at moderating content for an audience worth over two billion users. But how about deepfakes? Well, imagine someone created a pornographic deepfake video with your face and is now blackmailing you.

Meta won't be of much help. According to this Guardian report, Facebook is already failing to contain its "revenge porn" and "sextortion" problem. Heck, the company is now charging a fee of $12 to protect your account on priority. Now, tell me how many people you know that currently pay to use Facebook or Instagram?

The victim-culprit dilemma

Now, deepfakes are already harmful from a personal perspective. But when deployed with a coordinated psy-op intent, they can affect the whole nation. The easiest target? Elections. Slander an opposition leader with an unsavory clip? Check. Fake an audio clip alleging corruption? Check.

In a country where mainstream news channels often play clips from a video game and show it on prime-time shows as exclusive war footage, you can imagine how deep the risks are for the masses. But who would use deepfakes on such a massive scale to dupe an entire democracy?

Here's a clue. Remember the wrestlers vs government protest, in which a viral deepfaked image of the wrestlers was used to paint them as evil agenda-driven actors?

So, who stands to benefit from deepfakes, apart from petty criminals and online bad actors? The ruling party aka the Bhartiya Janata Party, led in the Parliament by none other than Modi ji.

All the way back in 2020, Vice exposed how the Delhi BJP IT Cell tapped a political communications firm named The Ideaz Factory to make deepfake videos that portray a positive image of the party and help with election campaigning. Those deepfaked videos featuring Bhojpuri star turned parliamentarian Manoj Tiwari actually worked their magic.  

They certainly have the money to spare on deepfakes. Of the 780 crores received by political parties in funding, BJP hogged nearly 80% of that sum. In a world where someone can create a full-length porn video of yours for as little as $30 by using just a 15-second clip lifted from the internet, imagine how much the tech is going to be abused to denigrate an election rival and win elections?

Fake news has already become an epidemic in India, and it seems like a lost battle already to imagine deepfakes won't be weaponized to deal more damage for electoral gains.

Deepfake trust deficit sits at the summit

But let's assume Instagram, Facebook, Twitter, and Reddit agree to proactively remove deepfakes. Will the union government let them do so? We've already seen how social media is bending over backward to silence the voice of critics while giving a place to the vilest content aligned with government interests.    

A collaborative investigation by The Reporters' Collective and AdWatch found how Facebook charged a smaller fee to the BJP for political ads. An Oxford research from 2019 found that ahead of the elections, 28% of news links shared on Facebook pages aligned with the BJP were categorized as "junk news."

In 2020, The Wall Street Journal reported how a top Facebook official took a stance against the application of the social media platform's hate speech regulations to at least one BJP politician (BJP) and various groups peddling the party's narrative. Heck, these platforms already seem hand-in-glove with the ruling party.

In a democracy where the government wants to be the arbiter of truth and self-fact-checker with the IT Rules (2021), it's hard to believe robust deepfake rules will ever be implemented. At least not before the elections are over.