A senior official in Narendra Modi’s government has warned that social media companies will be held accountable for AI-generated “deepfakes” posted on their platforms in compliance with “very clear and explicit rules” as India prepares for a general election this year.

Rajeev Chandrasekhar, minister of state for electronics and IT, said that India had “woken up earlier” to the danger posed by deepfakes than other countries because of the size of its online population. As many as 870mn people are connected to the internet while 600mn use social media out of a total population of 1.4bn.

“We are the world’s largest democracy [and] we are obviously deeply concerned about the impact of cross-border actors using disinformation, using misinformation, using deepfakes to cause problems in our democracy,” Chandrasekhar told the Financial Times. “We have been alert to this earlier than most countries because it impacts us in bad ways much more than smaller countries.”

The warning on fakes comes after Modi, who is seeking re-election to a third term in a parliamentary poll expected in April and May, broached the topic in recent remarks, and as India wields its regulatory clout over companies serving one of the planet’s largest populations of internet users.

Researchers have warned that deepfakes — images, video and audio created by cheap artificial intelligence tools that can convincingly recreate human beings — are a growing threat to democracies. Faked clips have already been used to influence politics and elections in the UK, Bangladesh, Nigeria, Sudan and Slovakia. 

Social media platforms meanwhile have drawn up rules to combat deepfakes. Meta, X and TikTok now require that deceptive manipulated media either be taken down or labelled. Meta and Google recently announced policies requiring campaigns to disclose if their political adverts have been digitally altered.

Some countries have been exploring clamping down on deepfakes in the context of pornography as well as politics, for example. In the US, for example, there are no federal laws explicitly governing the technology. However, multiple state lawmakers are pursuing legislation to tackle political deepfakes. 

New Delhi, in an advisory published on December 26 and sent to social media and messaging platforms active in India, including YouTube, X, WhatsApp, Telegram, Snap and local social network Koo, demanded that tech companies comply with Indian law on illegal content and make that clear in their terms of service and user agreements. 

India’s IT rules, drafted in 2021, also outlaw content that is deemed harmful to children, threatens national security or spreads misinformation, among other restrictions on free speech.

The directive warned platforms to “identify and remove misinformation which is patently false, untrue or misleading in nature and impersonates another person, including those created using deepfakes”. 

“Our approach to deepfakes after the prime minister drew attention to it is to tell the intermediaries, to draw their attention to the fact that the law of the land prohibits any user on their platform from hosting misinformation, including deepfakes, and casting an obligation on the platforms that, if any user does so, that user would be violating the law,” Chandrasekhar said.

He added: “By allowing the user to continue to have that content posted, the platform would be violating the law.”

He said that people made the mistake of conflating the US, where first amendment rights were “absolute and unconditional”, with the rest of the world, including India. 

“We are actually creating a form of tech regulation that is in between the US and Europe,” the official said. “The US leaves it completely to markets; Europe does it completely citizen-oriented; and we are basically saying ‘We love innovation, we will encourage innovation, but we want to protect our small businesses and users’.” 

In India’s most recent elections, held in 2014 and 2019, social media played an important role in the campaign waged and won by the ruling Bharatiya Janata party. Modi’s BJP is widely expected to win this year’s election in the face of a weak and divided opposition.

India is a sensitive market for the Silicon Valley companies as the Indian National Congress has accused platforms such as YouTube of “shadowbanning” some of their content, such as MP Rahul Gandhi’s speeches, while civil society groups alleged the tech companies bow too readily to Modi government takedown orders. In recent years, social platforms have been ordered to remove posts critical of for example the Indian government’s handling of the coronavirus pandemic and have blocked critics of Modi.

India is one of the biggest global markets for Meta’s platforms such as WhatsApp and Facebook, as well as Google and its YouTube video channel. Civil society groups accuse the Modi government of overzealousness in its policing of the internet, which they say is part of India’s broader crackdown on freedom of expression.

In a letter sent this month to Chandrasekhar and other officials responsible for setting government IT policy, the Internet Freedom Foundation, a non-governmental group, claimed that they could “disproportionately affect politically inconvenient or controversial speech, and potentially lead to arbitrary censorship”. 

Chandrasekhar rejected claims of government over-reach, saying that “nobody can argue that child sexual abuse material or deep fakes or paedophile or patent violating content is anybody’s infringement of anybody’s rights”.

“India is leading this charge, where platforms are being asked to take more and more responsibility and be more accountable for anything illegal that happens on their platform,” he said.