India Tightens AI Social Media Rules, Raising Censorship Concerns

India has announced stricter regulations governing the use of artificial intelligence (AI) on social media, aimed at curbing the spread of disinformation. Set to take effect on February 20, 2024, the new rules will significantly shorten the time platforms have to remove content deemed problematic, raising concerns about censorship and freedom of expression.

The regulations come as India grapples with a surge of AI-generated misinformation affecting its more than 1 billion internet users. Under the new framework, platforms such as Facebook, Instagram, and X will now have just three hours to comply with government takedown orders, a reduction from the previous 36-hour window. This change aims to prevent harmful posts from gaining traction.

Concerns Over Digital Freedoms

Rights groups are sounding alarm bells about the implications of these tighter regulations. Critics fear that the swift removal requirements might lead to excessive censorship and an erosion of digital freedoms. The Internet Freedom Foundation (IFF), a digital rights advocacy group, has characterized the new rules as potentially forcing platforms into a role of “rapid-fire censors.”

Apar Gupta, chief of the IFF, expressed concern that the compressed timeframe for social media takedown notices would hinder meaningful human review, stating, “The timelines are so tight that meaningful human review becomes structurally impossible at scale.” He further noted that this approach shifts control away from users, complicating grievance processes and appeals.

The Indian government, led by Prime Minister Narendra Modi, has previously faced accusations of curbing freedom of expression, particularly against activists and political opponents. The government’s policies have contributed to India’s decline in global press freedom rankings during Modi’s administration.

Automated Tools and Content Regulations

The new rules expand upon previous initiatives, including the launch of an online portal called Sahyog to automate the process of issuing takedown notices to platforms. The regulations now encompass content that is “created, generated, modified or altered through any computer resource,” with exceptions for routine editing.

Platforms are mandated to label synthetic or AI-manipulated media with permanent markings that cannot be removed or suppressed. Gupta highlighted the challenges of enforcing these labels, pointing out that metadata is often stripped when content undergoes editing or compression.

Digital rights activist Nikhil Pahwa described the regulations as “automated censorship,” emphasizing that most internet users remain uninformed about orders to delete their content. The rules also require platforms to utilize automated tools to prevent the dissemination of illegal content, including forged documents and sexually abusive material.

The Center for the Study of Organized Hate (CSOH) issued a report detailing that the regulations might lead platforms to err on the side of caution, potentially resulting in collateral censorship. The report defines synthetic data as information that appears to be real or is likely to be perceived as indistinguishable from actual events.

Gupta warned that the broad parameters for content takedown could sweep in satire, parody, and political commentary using realistic synthetic media, especially under risk-averse enforcement.

The implementation of these regulations reflects increasing global scrutiny over how social media platforms manage AI and the content generated through it. While the Indian government asserts that it is acting to protect citizens from irresponsible platform behavior, critics argue that the approach is fundamentally flawed and lacks consideration for the implications on free expression.