India Tightens AI Rules on Social Media Amid Disinformation Concerns
India has implemented stringent new regulations governing the use of artificial intelligence (AI) on social media platforms, a move aimed at curbing the proliferation of disinformation. However, these tightened rules have also raised concerns among digital rights advocates, who warn of potential censorship and a curtailment of online freedoms.
The new regulations, set to take effect on February 20, coincide with the conclusion of an international AI summit in New Delhi that brought together prominent figures from the global technology sector. A key aspect of these updated rules is a significant reduction in the timeframe platforms have to remove content deemed problematic by the government.
With a massive internet user base exceeding one billion individuals, India is actively confronting the challenge of AI-generated disinformation flooding its social media landscape. Platforms such as Instagram, Facebook, and X (formerly Twitter) will now have a mere three hours to comply with government takedown orders, a drastic reduction from the previous 36-hour window. This accelerated timeline is intended to prevent damaging or misleading posts from gaining widespread traction.
This stricter regulatory approach in the world’s most populous nation intensifies the pressure on social media giants. These companies are already facing mounting global scrutiny and public anxiety regarding the misuse of AI, including the spread of misinformation and the creation of sexually explicit imagery, particularly concerning children.
Despite the stated intentions, civil liberties organisations express apprehension that overly broad oversight of AI could inadvertently stifle freedom of speech. Critics point to a history under Prime Minister Narendra Modi’s administration where rights groups have accused the government of imposing restrictions on freedom of expression, often targeting activists and political opponents – allegations that the government has consistently denied. India’s standing in global press freedom rankings has also seen a decline during his tenure.
The Internet Freedom Foundation (IFF), a prominent digital rights group, has voiced its concerns, suggesting that the compressed takedown timelines will compel social media platforms to function as “rapid-fire censors.”
‘Automated Censorship’ and Shifting Responsibilities
In an effort to streamline the process of content moderation, India’s government last year introduced an online portal named Sahyog, which translates to “cooperate” in Hindi. This portal automates the issuance of takedown notices to platforms like X and Facebook.
The latest regulations expand the scope of these rules to encompass content that has been “created, generated, modified or altered through any computer resource.” The only exception is for material that undergoes routine or good-faith editing. A critical new requirement is that platforms must now affix clear and permanent labels to synthetic or AI-manipulated media, ensuring these markings cannot be removed or obscured.
Under these new rules, content deemed problematic could be removed almost instantaneously following a government notification. Apar Gupta, chief of the IFF, highlighted the impracticality of this compressed timeframe, stating that it makes “meaningful human review becomes structurally impossible at scale.” He further noted that this system decisively shifts control “away from users,” as “grievance processes and appeals operate on slower clocks.” A significant concern is that most internet users are not being informed when authorities order the deletion of their content.
Digital rights activist Nikhil Pahwa described the system as “automated censorship.” The rules also mandate that platforms deploy automated tools to proactively prevent the dissemination of illegal content, including forged documents and sexually abusive material. Pahwa expressed skepticism about the enforceability of unique identifiers for synthetic content, noting, “It’s impossible to do for infinite synthetic content being generated.”
Gupta echoed these doubts regarding the effectiveness of labelling. He pointed out that “metadata is routinely stripped when content is edited, compressed, screen-recorded, or cross-posted,” and that “detection is error-prone.”
Concerns Over Collateral Censorship and Online Hate
A report by the US-based Center for the Study of Organized Hate (CSOH), in collaboration with the IFF, has cautioned that these new laws “may encourage proactive monitoring of content which may lead to collateral censorship.” The report suggests that platforms, fearing penalties, will likely adopt an overly cautious approach to content moderation.
The regulations define synthetic data as information that “appears to be real” or is “likely to be perceived as indistinguishable from a natural person or real-world event.” Gupta elaborated that these changes shift the burden of responsibility “upstream” from individual users to the platforms themselves. He explained, “Users must declare if content is synthetic, and platforms must verify and label before publication.”
However, Gupta also warned that the criteria for takedown are broadly defined and open to interpretation. He expressed concern that “satire, parody, and political commentary using realistic synthetic media can get swept in, especially under risk-averse enforcement.”
Simultaneously, the widespread accessibility of AI tools has reportedly fuelled “a new wave of online hate,” as highlighted in the CSOH report. This wave is facilitated by photorealistic images, videos, and caricatures that reinforce and perpetuate harmful stereotypes.
A recent high-profile incident involved Elon Musk’s AI chatbot, Grok, which in January generated significant controversy when it was reportedly used to create millions of sexualised images of women and children by enabling users to manipulate existing online images of real individuals.
Pahwa acknowledged the government’s motivation, stating, “The government had to act because platforms are not behaving responsibly.” However, he concluded that “the rules are without thought,” suggesting a need for more nuanced and carefully considered regulatory frameworks.
