In Perspective | The only answer to India vs Big Tech on moderation

While India must defend itself from adversaries that will exploit social media for information warfare, it must, at the same time, protect free speech; all while social media companies have to do much more

Last week, news agency Reuters reported what it described as a “heated meeting” between Indian government officials and big technology companies Google, Facebook, and Twitter on fake news and content moderation.

For companies, they can start with upholding the Santa Clara principles — to which they are all signatories.  (Shutterstock)

Quoting unnamed sources, the news agency said India’s information and broadcasting ministry criticised the companies for dragging their feet on fake news, which forced India to point out content for takedown, which, in turn, opened the administration up to allegations of censorship.

During the meeting, one of the ideas — floated purportedly by a Google executive and shot down by the government representative — was for takedowns to become informal, which would have meant that these moved via unofficial channels.

The meeting, the report added, was a follow-up to the government’s efforts to block channels on Google’s YouTube for promoting alleged fake news and anti-India content. These, the government said, at a separate press briefing earlier in January, were part of an information warfare operations, meant to spread disinformation.

High stakes

Taking the government’s position at face value, the YouTube channels that were taken down for anti-India content were aimed at “undermining the democratic process of the upcoming elections in five states” and spreading “conspiracies about the death of chief of defence staff General Bipin Rawat”.

Indeed, there have been disinformation campaigns with seismic effects across the world — these have included deploying fake news factories that ultimately impacted the United States (US) elections in 2016 and the Brexit vote the same year, to sophisticated fake news operations that hired freelance journalists to target Left-wing voters in the US and Britain.

There is evidence of such operations targeting India as well. Cybersecurity analyst Pukhraj Singh summarised a particular Iranian operation, which was taken down by Twitter in 2018, as an attempt to weaponise prevailing political polarisation in Indian polity. Its outcome was not so much as domestic sabotage as it was a bid “to propagate a counter-narrative” in India.

In more recent times, India has alleged that Sikh separatist groups were employing similar strategies. A particularly recent allegation relates to a sexual assault incident in Delhi, where a woman was attacked and humiliated publicly in her neighbourhood. Days after the incident, there was a coordinated fake news campaign on Twitter saying that she had died by suicide, and her death was linked via hashtags — HT could verify these — as a larger plot to persecute the Sikh community to which she belongs. There had been no suicide.

A platform governance issue

All legitimate abuse of social media company products stems from two factors: Adversary sophistication in mounting such disinformation campaigns and the failure of policies and systems put in place by social media companies to check these.

The Indian administration’s reported anger at social media companies might make sense when instances of how some of these corporations have failed to deploy adequate protections are considered. Facebook whistleblowers have repeatedly flagged the lack of “classifiers” that will help the company detect hate speech and fake news in Indian languages, for instance.

Questions on the tools and policies are also inadequately answered in other instances. Most recently, this had to do with Opposition leader Rahul Gandhi’s Twitter account, which his Congress party said was under some sort of impression throttling — it was not banned, but there was a sort of shadow restriction that stopped the flow of followers to the Congress Member of Parliament (MP)'s Twitter account, the party alleged. An analysis by independent researchers found that this may indeed have happened, but the company has not adequately addressed reasons — whether Gandhi organically stopped receiving new followers or there were indeed algorithmic changes that played a role.

Censorship concerns

The missteps by these companies and the potential for abuse does not automatically mean that the government is right. There have been concerns raised at the government’s conduct and intention from within Parliament, the judiciary as well as civil society that have seen censorship risks in the rules and executive decisions the administration has taken.

Take, for instance, the judiciary. At least three high courts (HCs) — Kerala, Bombay, and Madras — have stayed the government’s information technology rules, 2021, which give the administration added powers to remove content online. The Kerala HC on July 9, 2021, said the new rules appear to give “excessive powers” to unreasonably and impermissibly” restrict freedom of speech.

Then there was a parliamentary committee report on internet shutdowns, which said there was no accountability mechanism in how authorities in the country ordered internet communications to be blacked out. The committee, with members from the Opposition and the ruling Bharatiya Janata Party (BJP), said the rules were lacking in clarity, there were inadequate safeguards and consultation and oversight was limited.

The report was “quite a strong indictment of the problem and the government’s approach to internet shutdowns, which is bringing India global infamy”, Raman Jit Singh Chima, Asia policy director at Access Now, told me at the time.

High stakes

The “global infamy” that Chima referred to is perhaps the implication that has been felt within the administration, if the Reuters report is anything to go by.

India was identified by an independent non-partisan American federal agency as among six countries where censorship as a policy and in practice posed a risk to US business.

The report found that “U.S. firms appear to have been increasingly affected in recent years by censorship in India” including, “in particular, a substantial increase in requests for content removal”.

The report in effect puts India in the same category as China, Russia, Indonesia, Vietnam and Turkey, and brings it to the spotlight of the American lawmakers. What effect this will have on diplomatic or trade relations remains to be seen, but there are several contexts to it.

The way forward

Concerns on either side are legitimate. India has the right to defend itself from adversaries that will look to exploit social media for information warfare and other cybercrimes. At the same time, it has to be held accountable for its constitutional obligation to protect free speech. And social media companies definitely need to do more.

For instance, whistleblowers Frances Haugen and Sophie Zhang have called on Facebook to release a Human Rights Impact Assessment that Facebook commissioned on its content moderation in India but is yet to release.

The answer, for both sides, lies in transparency.

For the administration, it can start with making takedown orders more detailed. Advocacy groups such as the Internet Freedom Foundation have routinely asked for this, pointing out that the exemption from needing to do so under Section 69A of the Information Technology Act enables secret censorship. This can become a start to address questions of overreach.

For companies, they can start with upholding the Santa Clara principles — to which they are all signatories. On content moderation, these principles advocate for companies to make public details of the type of systems and policies they deploy to take down content, and disclose any formal or informal working relationships with State actors. It also calls for companies to detail the State requests to take down or act against user content, and what action they took.

In Perspective takes a deep dive into current issues, the visible and invisible factors at play, and their implications for our future

The views expressed are personal