Introduction

The COVID-19 pandemic has shown how fast and wide information can spread: so fast, that the phenomenon was given the name, “infodemic”.^[1] For example, a documentary titled “Plandemic” featured the views of scientist, Dr. Judy Mikovits, who argued that the SARS-CoV-2 virus was “manufactured” and part of a scheme by pharmaceutical corporations to profit from selling vaccine. The documentary managed to rack up about 8 million views in 8 days.^[2] Such disinformation compromises society’s information-sharing ecosystem. In the time of a pandemic of massive proportions such as COVID-19, social media can be used as a tool of distrust to incite panic, confusion, and disharmony. The misuse of these platforms can have economic, psychological, and political impacts, both online and offline,^[3] and can lead to discrimination and even violence.^[4],^[5]

Behind the veil of protecting free speech, tech companies in India remain oblivious to such potential misuse. In the United States in early January 2021, platforms like Twitter provided a peek into their ability to counter disinformation, directing end-users to reliable sources, and suspending the account of former president Donald Trump, “due to the risk of further incitement of violence.”^[6] Many countries have initiated inquiries into the role played by these platforms in spreading extremist, hateful or fake content. Germany, Singapore, and France can now levy significant fines against platforms that fail to restrict illegal content after due process of notice and flagging. The United Kingdom (UK) is debating an Online Harms White Paper, while the European Commission has proposed two legislative initiatives—i.e., the Digital Services Act (DSA) and the Digital Markets Act (DMA) for the creation of regulatory mechanisms to counter online harms.

This paper seeks to provide an understanding of the Indian government’s approach to the massive regulatory challenges posed by social media platforms. A key question is whether the big corporations can be trusted to self-adjudicate while performing state-like functions to “reasonably” restrict speech. It is a relevant question, as Microsoft’s Digital Civility Index has found that Indians are most likely to encounter misinformation online.^[7] Other studies have found substantial growth in hate speech online.^[8]

This paper analyses key issues with India’s current regulatory framework to counter fake news and hate speech online, and draws lessons from the practices in other democracies. It argues that there is a need to modify the Indian approach by adopting a co-regulatory model. Governmental assistance through a penal and ethics code can complement the platform’s capacity to restrict online harms. Strictly independent obligations are recommended to advance the current content moderation practices. A multi-stakeholder collaborative approach is imperative to counter the threat of even worse infodemics.

Disinformation amid the Pandemic

The use of social media for peddling fake news and hate speech is not a new phenomenon. Before the pandemic, episodes of information dumping peaked during elections,^[9] socio-political movements,^[10] or to manipulate financial markets.^[11] Amidst the COVID-19 crisis, it has become apparent that widespread fake news can threaten public health.^[12],^[13],^[14] Public awareness is key in battling a health crisis. However, if the regulation of misinformation is concentrated in the hands of  platforms or government agencies, it becomes susceptible to perception-alteration tactics.^[15]

Early reportage of the pandemic in India tended to generalise Indian Muslims as wilful carriers of COVID-19.^[16],^[17] The biased reportage began after a religious gathering turned out to be a super-spreader. A countrywide search for the attendees began, and authorities discovered 4,291 affected; there were 27 deaths. Soon after, the hashtag ‘#CoronaJihad’ started trending in social media, in effect labelling a particular religious community as a collective danger to health and society.^[18] To be sure, the global cyber-space is infested with all types of hateful narratives that are successfully fostered within society’s digital communities.^[19]

At the same time, social media platforms serve a function in creating a more Covid-literate Indian society. Health and executive authorities use these platforms to spread awareness and direct end-users to reliable sources. Facebook’s MyGov Corona Hub, WHO Chatbot, and Corona Helpdesk Chatbot are a few examples of channels that have worked to counter misinformation and give access to first-hand awareness.

Facebook, for one, can be a highly powerful tool, with over 290 million users in India—its highest in the world.^[20] In recent times, however, various governments have begun scrutinising the platform for what they allege to be its lackadaisical approach to hate speech.^[21] In April 2020, Facebook  flagged 50 million  posts with warning labels; it  argued that once a content is flagged, 95 percent of end-users do not  access it.^[22] Fact-checking organisations are also working to counter fake news campaigns, including, in India—reports about purported “cures” against the COVID-19.^[23],^[24],^[25] According to a Reuters report, between January and March 2020, there was a 900-percent increase in fact-checks related to Covid-19.^[26] The same report indicates that a mere 20-percent of the total misleading content in that period had come from prominent public figures and enjoyed 69 percent of all engagement.

Social media platforms may have democratised the internet, but the same technology can create conflicts as it enables the proliferation of erroneous information at an unprecedented pace.^[27] In a 2017 tudy of the US, a team from the Massachusetts Institute of Technology  found that fake news spreads, “farther, faster, and deeper” on these platforms.^[28] The companies do not have adequate resources to quickly identify such content and remove them. The algorithms of these platforms work in such a manner that they record the user’s past interactions and fill their feed with their identified interests; this facilitates targeted advertisements, from where the platforms earn their incomes.^[29]

Of all the content in these platforms, those that are extremist, fake and populist are found to often garner high “interaction” numbers.^[30] Facebook, for example, took down 40 million misleading posts in March 2020 alone, and another 50 million the following month.^[31] For its part, Twitter challenged more than 1.5 million accounts from mid- to end-March.^[32] However, Network Contagion Research Institute (NCRI) highlighted the role of smaller online communities and groups, that have become active avenues for targeted divisive content. While studying the “China-led bioweapon controversy”, NCRI found profound anti-Asian and Sinophobic conspiracy theories on a small website, often being the source of disinformation and propaganda on other platforms like Reddit and Twitter.^[33]

The question therefore, is if these platforms are plagued with manipulative and unethical content, can they still democratise? In theory, the principle invoking the so-called “marketplace of ideas” is a bedrock of free speech laws; it presumes that “truth” would prevail in a level playing field.  In the marketplace, however, of platform-based ideas, the theory fails—^[34] it seems social media is neither equal nor fair. A platform’s design to maximise financial gains through data monetisation techniques can overwhelm “truth” with inbuilt susceptibility to sensationalist, viral, curated campaigns. Problematic speech is heightened due to the asymmetry of information and polarisation online.^[35] To operationalise a model with necessary safeguards, the Indian approach must depart from excessive criminalisation or take-downs and instead adopt a holistic approach.

India’s Regulatory Framework: An Overview

Fake News

There is inadequate regulation of fake news under Indian law.^[a] Due to the various types of fake news, their motivations, and the ways they are shared, the regulatory challenge is daunting. To combat fake news, the first imperative is to identify the different forms: ‘misinformation’ is the inadvertent sharing of false content, while ‘disinformation’ is deliberate sharing with an intent to deceive. Its sub-types, according to Claire Warlde of US-based First Draft, are misleading content; imposter content; fabricated content; false connection; false context; manipulated content; and satire or parody.^[36] The misuse of platforms is also aided by so-called “flooding tactics”, enabled through inorganic sharing or misusing emerging tools, like Deepfakes.^[b] In a case study conducted by Blackbird.AI, a US-based company, a dangerous propaganda campaign was unveiled wherein 99.4 percent of 927,908 tweets in 47 different languages were found to have been inorganically disseminated in early February 2020.^[37] A widely followed online religious leader and his followers waged a hashtag campaign, #NoMeat_NoCoronavirus, which sought to exploit religious bias through bots or impersonating user profiles.

Fake news thrives on dissemination through surplus or deficit information models. Under the surplus model, if enough users share the same information, it validates itself by a sheer numerical advantage, including when the gatekeepers of information (like journalists or politicians) validate it. A deficit information model suffers from a shortage of correct information, showcasing a lack of successful transmission between government, media, experts, and citizens. Information fails to trickle down and is manipulated en-route. Its impact is enhanced due to lack of access to correct information, limited prominence of fact-checking mediums, overwhelming nature, or the user’s inability to comprehend its consequence. It also garners authenticity and virality through self-authentication,^[38] and stylistic elements (like graphic content, alarmism, and imitation of authority).^[39],^[40]

The Indian Ministry of Electronics and Information Technology (MeitY) has recognised the potential for misuse of platforms and even broadly defined ‘disinformation’.^[c][41] However, the term is yet to be adopted under the IT Act or any provisions of the penal code. Section 505(1)(b) of the Indian Penal Code or Section 54 of Disaster Management Act, 2005, both provide broad recourse against cases which have severe consequences on public wellbeing; they are shorthanded, however, against the rapid pace of social media.^[42] These regulations also lack precedent or uniform application against multiple types of fake news.

Hate Speech  

Absolute free speech laws that protect against any type of censorship inadvertently render protection to hate speech as well. In India, hate speech is not profusely restricted, it remains undefined with appropriate IT Act provisions or a regulatory mechanism for online content. Absent appropriate codes or regulations for intermediaries,^[d] those who tend to have a louder voice—such as politicians or celebrities—can harness this capacity to incite anger or divide communities without being threatened by any form of liability.

India’s multiple laws on sedition, public order, enmity between groups, and decency and morality, broadly form the country’s jurisprudence on what is known as “hate speech”, without using the term itself. Following the unconstitutionality of Section 66A of the IT Act,^[e] no provision under the IT Act currently aims to curtail either online or offline ‘Hate Speech’. The most employed sections 153A and 295A of the Indian Penal Code (IPC) are also inadequate to deal with the barrage of online hate content. The Parliamentary Standing Committee has recommended changes to the IT Act by incorporating the essence of the Section 153A.^[43] The report also suggests stricter penalties than prescribed under Section 153A due to the faster and wider spread of information in online spaces. It advocates criminalising “innocent forwards”, for example, with the same strictness as the originator of the content.

This approach has the potential to create a “chilling effect”: the overcriminalisation of online speech.^[f] These provisions remain vaguely worded, and lack consistent interpretation of existing framework across courts. They also lead to the filing of minor cases that only overburden the country’s courts and fail to act as redressal.^[44] Unfortunately, unconstitutional provisions like Section 66A of IT Act continue to be invoked for prosecuting individuals. This provision is often misused due to its cognizable nature (allows arrest without a warrant) and highlights a signal failure between branches of the government.^[45]

To address the proliferation of hate speech on social media, criminal law should not be the first resort, but the last. The promotion of non-regulatory tools—i.e., counter speech, fact-checking, and digital education—is the imperative.

The Social Media Regulatory Dilemma

Both government authorities and social media platforms alike, have been criticised for their failure to secure data and effectively regulate content. Many platforms, experts, and politicians have welcomed a government-led moderation of illicit content, with ample checks and balances against arbitrary imposition.^[46],^[47] Human right groups and activists express skepticism against allowing any avenue for governmental intervention through either the arbitrary imposition of bans, content moderation, or internet shutdowns.^[48] Another paradigm champions the principle of “self-regulation”—where the platform itself adjudicates on their user-policy and community guidelines. Self-regulation has largely been ineffective in preventing abuse of the platform and has garnered criticism in various democracies.^[49],^[50],^[51],^[52] Indeed, both government-led moderation and self-regulation models have been operational worldwide and in India since the inception of social media. For example, more than 17,444 websites were blocked for promoting obscene content until 2019 by the IT Ministry^[53] and similarly, Twitter took down 636,248 accounts in 2015-16 alone for disseminating extremist content.^[54]

The difficult question concerning hate speech or fake news legislation pertains to the existing ethical-legal gap, the executive response departing from conservative understanding of online spaces and data. While disruptive technologies are evolving at a faster rate, the regulations fail to address gaps to deter unethical behaviour. The platforms alone are not equipped to oversee the task for a remodelled approach to counter manipulation and hate speech. Due to the overarching jurisdictional nature of these acts and easy multiplication, taking down content is not a silver bullet in countering hate speech and fake news.

While intensifying fact-checking and taking down inflammatory and fake content is a necessity, complex content-driven issues have emerged. For example,^[55] Facebook in India has been accused of “ideological bias” by both Left- and Right-wing groups.^[56] The Union Minister for Information Technology has called it “inherently biased” against people who support right-of-centre ideology and a “latest tool to stoke internal divisions and social disturbances.”^[57] Even as Facebook merely repeats its response—that it has an intermediary role as a free speech enabler—it has come under scrutiny again, most recently by the Parliamentary standing committee.^[58] Moreover, a group of Facebook employees who identify themselves as Muslim@ wrote an open letter to the Facebook leadership in August 2020 demanding greater transparency in taking down content; they also questioned why anti-Muslim and hateful content continue to find space on their platform.^[59]

The lack of accountability and transparency calls for a rethinking of social media platforms’ role and structure in order to counter their misuse.

• Structure:

The overregulation vs. under-regulation debate tends to overshadow the deeper and more inherent structural problems in the tech platforms themselves. The platform structure is driven by exploiting the disparities of wealth and power, as algorithms reward virality and interactions for monetary gains, even though they might be “divisive, misleading or false”.^[60] Platforms are also known to amplify certain types of users and content over others.^[61],^[62] Platforms decentralise free speech, but “special” megaphones are provided to sensationalist ideologies or popular content. Its algorithmic nature creates and perpetuates an information divide, alienating communities with different subscriptions through echo-chambers and information silos. This has become obvious with the platform’s incentive structure, which is driven by monetisation of user data, advertisement money, and constant engagement. For example, a few popular YouTube channels that earlier achieved “Creator Award” were inciting violence including rape but suffered less takedowns.^[63] Platforms conveniently hide behind the garb of free speech enablers, with little responsibility, if at all.^[64] Even as xenophobia, communalisation and racism have long existed in the real world, the susceptibility of social media platforms to misuse has magnified such ill-speech at a faster pace.

In India, social media platforms are not liable under any rules or regulations. They function under a regulatory vacuum and are not bound by any industry regulatory standards for the functions they dispense. None of existing news agency regulations, consumer protection laws, data privacy or other traditional sectoral regulations apply to these intermediaries. The bigger the platform, the less resources are available for user redressal against online harms. Functions like reporting or flagging:  inflammatory and fake content are opaque, ineffective and inconsistent.^[65] In some cases, the lack of competitors offers no incentive to provide better user protection, with majority of the targeted population already operating on the platform. Yet, the platforms continue to fail to recognise their public utility functions. In various countries, companies like Facebook, Google, and Twitter have succumbed to governmental crackdown, acknowledging the lack of preparedness and instituting duties to protect users.^[66]

• Role:

The diminishing optimism amongst policymakers that tech companies can self-regulate has raised concerns on the unintended yet significant consequences that accrue from the current understanding of a platform’s role. Under Section 79 of the IT Act, a government order can direct an intermediary to take down problematic content and not face any liability as publisher. Other than complying with state-determined policy, platforms also moderate harmful and unethical use through their own policies.^[67]

Many argue that social media’s role has evolved, as these platforms perform more than an intermediary role as content host.^[68] Through the “community guidelines” that they set for their users, they are in effect performing the state-like legislative function of outlining rules for netizens. Similarly, they conduct state-like content moderation functions by taking down objectionable content.

The need for safeguards against problematic speech remains imperative and the role played by intermediaries must advance in tandem while revising its fiscal incentive structure and its public utility role.

Best Practices in Regulation: Key Lessons

India can look to other countries for lessons on regulating social media platforms. Countries like Germany, the US, UK, Singapore, and Russia, for example, have put in place specific accountability standards assigned to these platforms. The approaches taken by these countries, however, are varied.

Table 1. Regulatory standards for illicit content

Source: Author’s own, using various open sources.

Australia

Australia’s response to the proliferation of fake news is focused on countering foreign interference during elections. This was set up in the backdrop of growing concerns of potential Chinese interference and 2016 Russian interference in US elections. A task force was created—called the Electoral Integrity Assurance Task Force and led by the Home Affairs Department—to strengthen cybersecurity.^[71] Two key frameworks are under discussion—the Fake News code^[72] and the News Media Bargaining Code. The News Media Bargaining Code is highly controversial from the perspective of copyright and competition law, as it forces big companies such as Google and Facebook to realign market power by sharing multi-million-dollar revenues with newspaper and media companies for using their news content, to aid their financial futures.^[73] Among the strictest penal provisions (in terms of fines) against extremist content—the Abhorrent Violent Material Act, 2019—make tech executives liable for imprisonment up to three years and fines for the platform can run up to 10 percent of a company’s global turnover.^[74]

France

French President Emmanuel Macron stated in November 2018, “if we do not regulate Internet, there is the risk that the foundations of democracy will be shaken.”^[75] France applies broadcasting standards for TV and radio stations to social media through the Higher Audiovisual Council (CSA), the independent media regulator. It aims to assure broadcasting communication freedom in the country and has been tasked with the responsibility of enforcing state policies on fake news and hate speech. The French government has identified terrorist content and child pornography as non-negotiables, demanding platforms to take down such content within one hour of being notified. France has initiated a crackdown on electoral misinformation to avoid foreign interference by empowering the CSA to unilaterally revoke broadcasting rights of news and radio outlets that function “under the control or influence of a foreign state” to disseminate misinformation.^[76] At the same time, however, the law has faced criticism for granting executive and administrative powers to CSA.^[77]

France has had a strict Hate Speech law in effect since 2005 and has passed a stricter legislation for online application against hateful speech (Lutte contre la haine sur internet- An Act to Combat Hateful Content on the Internet- Avia law)—it requires social media platforms to take down objectionable content within 24 hours of being notified. Once a political party, public authority, or an individual has filed a specific request, a judge is authorised to act “proportionally” but “with any means” to halt the dissemination of the fake news in question. The request is to be acted upon within 48 hours and a similar timeframe is applicable during an appellate request. France has also demanded transparency reports about the algorithmic functions of social media platforms, as well as financial transparency for sponsored content by disclosing the identity of promoters. The CSA is also tasked with publishing reports on data, mechanisms adopted, and their effectiveness.

Under France’s Digital Republic Act, “a decision taken based on an algorithmic treatment” by public sector bodies falls under a transparency rule called “right to explainability”. The citizens upon request can seek an explanation on the rules and its “principal characteristics.”^[78]

Germany

Germany’s NetzDG has jurisdiction over high-user platforms; it can order these companies to take down illegal content or else face significant financial liabilities.^[79] Human rights activists, however, are concerned about “over-removal” and privatised enforcement. The group, Human Rights Watch, has warned that the current regulatory framework “turns private companies into overzealous censors to avoid steep fines, leaving users with no judicial oversight or right to appeal.”^[80] Indeed, while the law was aimed at incentivising the quick take-down of illegal content, it has led to censorship. To begin with, the definition of “unlawful content” remains highly debated in instances of blasphemy and hate speech; “defamation” and “insult” are vaguely defined as well.^[81] Another counterproductive outcome can be explained as the ‘Streisand Effect’:^[g] prominent cases of deletion may fuel anti-government sentiments by garnering more publicity for the deleted content.^[82] Moreover, any extremist or banned user can still easily migrate to smaller platforms who are not liable under NetzDG.^[83]

Singapore

Singapore’s POFMA Act is arguably one of the most comprehensive misinformation laws globally. It is widely criticised by human rights groups and free speech advocates due to its vulnerability to governmental interference.^[84],^[85] POFMA covers “false facts” or “misleading statements”, but not satire, parody, opinions, and criticisms. Under POFMA, various directions can be issued to platforms and users before dispensing fines and imprisonment. These directions include the following:^[86] (i) Correction Direction: the party who communicated the falsehood is tasked to issue a public notice stating the false nature of information. The notice is published in a newspaper or a designated area on the platform.  (ii) Stop Communication Direction: the publisher is directed to ensure the information is no longer available for proliferation, with the direction being published in a government gazette. (iii) Targeted Correction Direction: the platform is directed to send a correction notice to all end-users who accessed the falsehood. (iii) Disabling Direction: the internet service provider is directed to disable access to the falsehood.

Another key aspect of the law is gross penalisation of inorganic proliferation methods like the use of bots.^[87] It is strictly prohibited to allow another person to manipulate or create bots to spread misinformation online, and violators face imprisonment of up to 10 years. Amendments under the Protection from Harassment Act (POHA) are under discussion to provide recourse to victims of harassment due to falsehoods.

United Kingdom

The UK released its Online Harms White Paper^[88] in April 2019, which is based on installing an overarching principle of “duty of care” towards end-users. The proposed Online Harms Paper is not limited to social media platforms; it extends to file-sharing sites, discussion forums, and e-commerce websites–which are mandated to take “responsible steps” in the direction of user safety, transparency and tackling harms. It calls attention to the need to empower users by developing their digital skills for countering misinformation and other forms of online harm. An independent regulator will be appointed with the power to issue substantial fines for social media platforms and their senior members. The paper dictates a risk-based approach, which aims proportionate regulatory action to counter harms with greatest impact on individuals. It is overwhelmingly positive and has initiated a discourse to prioritise before hurried legislative battles. However, the scope of the paper remains wide enough to cover various legal and ethical norms. At present, the National Security Communications Unit is tasked to combat disinformation campaigns by state actors and others during elections.^[89]

United States

The US guarantees absolute freedom of speech and has no legislation on hate speech. Through various judicial pronouncements, courts have consistently argued that even hate speech is protected under the First Amendment.^[90] The Russian interference during the 2016 presidential elections served as a turning point in the threat perception towards misinformation and manipulative sponsored content.^[91] In the 2020 presidential elections, big platforms like Instagram, Facebook, and Twitter took proactive steps to flag questionable content and direct end-users to reliable sources.

The Honest Ads Bill, introduced in 2017, seeks to extend the purview of ad regulations for broadcast TV and radio on social media platforms.^[92] It expands source disclosure requirements as “public communications” or “electioneering communications”, demanding source disclosures and the information contained within ads. The bill also requires social media platforms to maintain publicly available records about qualified political advertisements. Various state governments across the US have initiated compulsory digital and media literacy initiatives for schools. The security implications arising from deepfakes has also been highlighted by members of both the Republican and Democratic parties.^[93] Section 230 of the Communications Decency Act (1996) provides the intermediaries with broad immunity: they are not liable to take down any offensive content that upholds free speech. Only in cases of copyright infringement under Section 512 of the Digital Millennium Copyright Act (1998), can the state employ a “safe harbour” approach where it provides the conventional “notice and takedown” method.

Framing India’s Approach

In different parts of the world, disinformation and hate speech related to the COVID-19 pandemic have been addressed with specific responses. The pandemic serves as another case study for Indian jurisprudence, and can assist in reducing the ethical-legal gap. Online speech regulation necessitates improved understanding of online harms, their ill-effects on a democracy to sharpen the local response.

The Indian response must be driven by four guiding principles: (1) Accountability and transparency over decision-making by tech platforms, state and non-state actors; (2) Ensure consistency and collective will by encouraging inclusive stakeholder engagement for all decision-making processes; (3) Respect human rights standards and habituate humane application of tech. Incentivise innovative adoption of redesigned tech products that pre-empt and provide safeguards from online harms; (4) Legal certainty for consistent application and execution of duties and rights of stakeholders.

The lessons from the global best practices discussed in this paper should serve as basic elements for India’s regulatory mechanism: (i) Institute an independent regulator to oversee compliance with fake news and hate speech codes that will be adopted; (ii) Proportional, necessary and minimal interventions from the government and platforms with effective and consistent application of their duties; (iii) An inclusive and ethical Code of Conduct developed in consultation with all stakeholders to realign platform’s fiscal-driven-incentives with public interest; (iv) Democratic application of penal and non-penal standards of existing laws; (v) Periodic review policies to improve effectiveness; (vi) Encourage transparency by commissioning open-source research with periodic reports from regulator, platforms, civil society organisations and academia; (vii) Avoid creating any barriers or strengthening any dominant positions by large incumbents; (viii) Promote digital education initiatives and workshops to acquire necessary skills from a young age; (ix) Redressal and appellate mechanisms to provide support to any wrongful application of standards, take-downs or breach. 

Specific Recommendations

In any strategic intervention against misinformation and hate speech, the tech platforms are bound to play the biggest role. There should be continuous collaborative engagements within the industry, along with state and non-state actors. While the creation of charters or codes that define each stakeholder’s duties and rights will be a lengthy process, a pre-emptive plan cannot be delayed further. This can enable the creation of voluntary multi-platform and multi-stakeholder initiatives. The Code of ethics and voluntary audits are another welcome by-product of these collaborative measures. Issue-specific methods of advertisement rules for transparency and media guidelines or ethical codes also aim to strengthen industry standards. Some shared responsibilities between the stakeholders have already been outlined but limited action has been taken to counter online harms.^[94] Platforms have deployed minimal resources to take down blatantly illegal content, as they lack real-time local responders who are well-versed in Indian languages.^[95] Even their community guidelines are globally uniform and limited due to implementational and definitional challenges locally. Therefore, the government and the tech platforms should complement other information gatekeepers like media and politicians.

First, the government must speed up defining and paving a way for consensus towards a legal framework against problematic social media content.

• Definitions:

The Indian challenge to garner consensus and counter ‘hate speech’ and ‘fake news’ extends to their understanding in real/offline world. Both remain undefined under any domestic legal mandate, including the IT Act. By building consensus on key elements, the legislature can assist the platforms in initiating a countermovement by consistent interpretation and implementation of the law. It will also discourage platforms from adjudicating on what is acceptable speech and avoid faulty implementation. The Hate Speech Law Commission report, which suggests that the scope of regulation should not be limited to “incitement of violence” but also prohibit advocacy of hate; and incitement to hostility or discrimination.^[96] To determine penal implications for ‘disinformation’ and ‘hate speech’, ‘impact’ and ‘intention’ are key elements to discourage its proliferation. However, the process of defining and introducing penal provisions must avoid ambiguous terminologies. Cohesive definitions can complement adoption of voluntary codes amongst platforms, as well as update media code and Representation of the People Act (1951) against information manipulation during elections. Potential overcriminalisation must be prevented, the legislature can identify and agree on key elements to facilitate consensus building and build safety nets around ethical codes. Criminal law should not be the first response but the last resort when state or court intervention is imperative.

• Legislative Framework

A limited but necessary legislative support could help generate consensus with a much more effective regulatory mechanism.^[97] Countries like France and Germany have pre-defined the limited scope of government role to avoid any arbitrary intervention and have even appointed independent regulators or independent judicial members to dispense moderation objectives. Fake News and Hate Speech codes must be founded, and authorities must explore the scope of an independent regulator. The regulator’s objectives should be to assist the application of government policies and serve as a forum for redressal against the platforms for arbitrary takedowns. The regulating body by design must not be subjected to governmental or platform discretion that can have the potential to be selective, arbitrary, non-transparent, and utilised to smother dissent or the genuine exercise of freedom of speech.  The appointment of personnel in the body must also be inclusive, from a wide-ranging pool of judiciary, civil society, free speech activists, and technical experts. The code must curate appropriate guidelines to define platform duties, rights, and issue take-down notices through the regulator.

The code must differentiate between infrastructure information intermediaries (such as Internet Service Providers) and content information intermediaries that host content (such as platforms). A set of content-neutral regulations must apply to both intermediaries with a duty-of-care imposed on content information intermediaries. The under-discussion Intermediary guidelines could reflect recommendations that will define the role of the regulator, but an overarching code/framework is needed to restrict fake news and hate speech. The Code must foster transparency as a pillar and enable publication of takedown notices requested by the regulator. The Code must also provide recourse for due process and charge reasonable fines upon breach by any platform or individuals, and guarantee safeguards against potential abuse of notice periods, proper procedure, and wrongful takedowns.

The regulatory body can also assist in formulating and strengthening industry standards for the under-discussion Information Trust Alliance (ITA). ITA envisions a collaborative effort to bring all platforms to assist in an inter-platform exchange, it advises effectiveness through knowledge sharing. ITA can provide invaluable contribution in creating guidelines that highlight local problems, bottlenecks, effectiveness in diverse regional languages, study local trends, revaluate right to explanation and share data in a transparent manner.

Sponsored content and political ads should also be compulsorily fact-checked while maintaining directories of promoters, amount paid, and source. The use of inorganic amplification methods like bots to propagate hate agendas must be charged with fines in case of severe social impact. Gatekeepers of information like media houses and politicians owe a higher ‘duty-of-care’ as they yield a significant impact on local perceptions. Penal fines that are proportionate and consistent against repeat offenders must be employed.

However, learning from Singapore’s example, inconsistent and vaguely worded law with strict government-centric censorship will be a step in the opposite direction. Singapore’s fake news law is unconstitutional by Indian standards and criminalisation-heavy. Indian policy and law should instead reflect the necessary behavioural change that is required to address the awareness with a modified platform-incentive model. Internet shutdowns are not feasible measures as they have economic and social implications, worsening enmity and lack of correct information.

• Social Media Platforms

To effectively safeguard against and mitigate the impacts of ill-speech, countries should have better fact-checkers and authorities committed to respond to public interest. Real-time takedowns can have a positive impact as it significantly develops resistance and reduces impact of misinformation to percolate into local, deeper, and different forms. However, taking down problematic content is not an absolute safeguard in itself as the nature of platforms dictate that time is not the only determinant when problematic speech is published online. For example, a tweet might be taken down but still may have a direct and viral effect through its screenshots on different platforms. Ill-speech’s omnipresence and negative impact remain an issue but the lack of awareness amongst end-users showcases disengagement with the primary stakeholder.

Other than adopting voluntary responsible measures, this paper suggests a four-step model to be implemented by social media platforms to counter problematic content online.

• Step 1- Identification: Identify fake news, extremist content, or hate speech according to the definitions or elements. While upholding anonymity, the platform must flag it in a specific manner which communicates its problematic/unreliable nature to the end-user.
• Step 2- Disallow Proliferation: While the content may continue to exist online, it should not only be flagged but platforms should disable any type of proliferation further, which includes content’s algorithmic prioritisation it may usually employ. Any blatantly problematic content should not be promoted for interactions—i.e. such content should not be available on user feeds. Disable or provide safeguards for like, share, retweet, upvote, or other interaction methods for the same.
• Step 3- Issue interaction warnings: One of the most important criticisms against platforms is related to their collection of data. However, this issue can be employed to initiate a vital process of digital education and awareness. Since platforms employ interaction data, they must issue warnings to all end-users who have encountered problematic content before it was flagged or identified. All end-users who shared or promoted such content must be sent personal notifications on the respective platforms, about the problematic nature of the content. Directing them to more credible sources and encourage corrections on smaller groups. Similarly, the publishing end-user must be provided with necessary reasons for flagging or taking down the published content. They, along with other interactors, should be encouraged to fact-check before sharing published content. This can help in disseminating digital education and necessary online skills to empower the end-users themselves.
• Step 4- Provide a better recourse mechanism: In terms of reporting fake and hateful content, platforms should be user-friendly with timely action and response. This would require a wide expansion of resources employed at the intermediary’s offices. Recourse against wrongful takedowns should be formalised and direct end-users to such mechanism if their content is taken down.

Publishing and providing access to relevant data is key for studying local trends in speech manipulations. Through transparency reports, data should be made publicly available for better preparedness and constant engagement with all stakeholders to improve effectiveness. Platforms should also use disruptive technologies like artificial intelligence, which have been touted to industrialise mass-level identification of such content. They can also assist political institutions and media houses to countercheck information before posting it. For example, Facebook had introduced a Corona Helpdesk Chatbot in messenger to counter COVID-19 related misinformation online.^[98] Although AI has its limitations as well, for example, it has been unable to identify AI-generated misinformation itself and its use by platforms to flag problematic content, has often flagged non-spam or news content.^[99] This highlights technological gaps while employing AI. Various fact-checking initiatives exist outside social media platforms but their use and inculcation on platforms can be made user-friendly through building partnerships. Lastly, platforms must be incentivised to adopt innovative redesigned tech products which prioritise humane use, prevention of harm, and provide safeguards from harms that may accrue through their misuse.

• The end-user

The end-user who is exposed to problematic speech is the most important, yet, disengaged stakeholder. As social media platforms aim to ensure their risk-averseness, they must be a part of the safeguarding process. Inculcating anti-fake and anti-hate behaviour is key in empowering individual entities to discredit and stifle such speech as the first response. Counter-speech means to engage directly against the negative speech using positive expression. In online discourse, multiple end-users already assist towards “toning down the rhetoric”^[100] by providing clarifications to dubious claims, directly engaging with facts to counter hateful messages, call out fake reportage, employ humour or dissent through memes and caricatures. Most importantly, counter-speech is also a decentralised method for engaging with ill-speech, it reduces negative impact and counters the intention behind it. Radical change in user behaviour incentives is imperative to promote counter speech, disincentivise blind sharing, inculcate fact-checking, and calling-out disfavoured speech. There is no countermeasure against blind sharing but to effectively reduce its impact, it is necessary to empower individuals with easy access mechanisms to halt negative proliferation. A red flag should be easy to raise against extremist speech while imparting understanding against different types of ill-intentioned speech policies, potential impact, and its omnipresence.

Influential political, religious, or social leaders often misuse the significant online traffic while disseminating information. Trends suggest they enjoy widest amplification powers. For counter speech, incentivise best-placed “credible messengers” like politicians, media, celebrities, experts, and other individuals (with a lot of followers or authenticating blue ticks) to educate against fake news. The impact of their sharing ecosystems has a higher impact on end-users. For example, Facebook’s CEO Mark Zuckerberg justified the need to take down objectionable speech employed by Indian politician Kapil Mishra^[h] that intentionally “incites violence”.^[101] Hate speech originating from a politician has even more detrimental impact in a volatile environment. Political institutions must recognise and impose a ‘duty-of-care’ amongst politicians for humane and ethical use of these platforms.

• Media:

In countries like the US and France, some broadcast media regulations are made applicable to advertisement and sponsored content through declaration of sources to avoid foreign intervention in local politics. In India, independent bodies like the Editor’s Guild and Press Trust, should reset industrial standards and wield best practices to discourage competitors from providing space to fake, propaganda, and hateful narratives. A fine structure against repeat offenders both online and in broadcast media should be applicable, with terms of suspension and even revocation of broadcasting license in extreme cases.

Unlike social media platforms, in case of offline broadcast media a key feature of being able to compartmentalise news through subscription is absent. In a recent development, Supreme Court issued a pre-broadcast injunction against Sudarshan News for communally charged and inciting news show titled ‘UPSC Jihad’. SC expressed its concern of an ineffective regulatory model for broadcast media and questioned lapses by Press Trust of India to successfully restrict hate. The overlapping nature of broadcast and digital media also resulted in the hashtag ‘UPSCJihad’ trending on Twitter. Such lapses can be pre-empted within a co-regulatory platform model, that ensures not just compliance with ethical standards but is complemented by an incentive structure to disseminate facts-based reportage. The right to reply should not be absent in case of sensationalised speech.^[102]

Conclusion

The proliferation of hateful and harmful information online has become so widespread, that the United Nations, through Secretary-General António Guterres has underlined the need to counter what he called a “tsunami”.^[103] Guterres has also called attention to the scapegoating and “disproportionate effects” on vulnerable groups: they lose access to healthcare, and become victims of targeted violence, rising stigmatisation, and heavy-handed security responses. Proportionate restrictions and guidelines have not developed in tandem with the expansion of the power of the social media platforms. Users are exposed to problematic content without any warning, awareness, and skills to counter what they are against. A social media platform’s primary incentives remain their massive user base, increased sharing and connectivity, and profits.

The evolving nature of online harm necessitates an appropriate response from the regulatory bodies. Additionally, the dissimilar nature of the pandemic, compounded by the weaponisation of information-sharing models, benefit few and negatively affect large populations.

Intervention in this regard is necessary. However, any restriction cannot be vaguely or hastily drafted to allow selective and arbitrary application by either the tech companies or government authorities. A balance must be found in this regard, defining the roles of various stakeholders in a co-regulatory model. Principles of proportionality, accountability and transparency must serve as cornerstones of any speech regulation. The pandemic case study offers another learning curve for collaborative countermeasures.

Hate speech is provocative and divisive, and in extreme scenarios where it has remained unchecked, has been responsible for terrorism and genocide. With newer tools to weaponise and sensationalise enmity, it must not be protected under the realm of free speech doctrine. Similarly, misinformation (“fake news”) also has the potential to affect human safety and public health, and instigate violence. If fake news and hate speech continue to proliferate at the current rate, they pose threats to the democratic ecosystem. India must work to devise an all-stakeholder model to counter the weaponisation of online content, before it further widens societal faultlines.

About the Author

Archit Lohani is a Research Assistant at ORF.

Endnotes

• Cyber Security
• Law and Justice
• India
• cyber security
• hate speech
• TECH AND MEDIA

The views expressed above belong to the author(s). ORF research and analyses now available on Telegram! Click here to access our curated content — blogs, longforms and interviews.

PREV NEXT