Resurrecting the marketplace of ideas

The article by Arindrajit Basu was published in Hindu Businessline on February 19, 2019.

A century after the ‘marketplace of ideas’ first found its way into a US Supreme Court judgment through the dissenting opinion of Justice Oliver Wendell Holmes Jr (Abrams v United States, 1919), the oft-cited rationale for free speech is arguably under siege.

The increasing quantity and range of online speech hosted by internet platforms coupled with the shock waves sent by revelations of rampant abuse through the spread of misinformation has lead to a growing inclination among governments across the globe to demand more aggressive intervention by internet platforms in filtering the content they host.

Rule 3(9) of the Draft of the Information Technology [Intermediary Guidelines (Amendment) Rules] 2018 released by the Ministry of Electronics and Information Technology (MeiTy) last December follows the interventionist regulatory footsteps of countries like Germany and France by mandating that platforms use “automated tools or appropriate mechanisms, with appropriate controls, for proactively identifying and removing or disabling public access to unlawful information or content.”

Like its global counterparts, this rule, which serves as a pre-condition for granting immunity to the intermediary from legal claims arising out of user-generated communications, might not only have an undue ‘chilling effect’ on free speech but is also a thoroughly uncooked policy intervention.

Censorship by proxy

Rule 3(9) and its global counterparts might not be in line with the guarantees enmeshed in the right to freedom of speech and expression for three reasons. First, the vague wording of the law and the abstruse guidelines for implementation do not provide clarity, accessibility and predictability — which are key requirements for any law restricting free speech .The NetzDG-the German law, aimed at combating agitation and fake news, has attracted immense criticism from civil society activists and the UN Special Rapporteur David Kaye on similar grounds.

Second, as proved by multiple empirical studies across the globe, including one conducted by CIS on the Indian context, it is likely that legal requirements mandating that private sector actors make determinations on content restrictions can lead to over-compliance as the intermediary would be incentivised to err on the side of removal to avoid expensive litigation.

Finally, by shifting the burden of determining and removing ‘unlawful’ content onto a private actor, the state is effectively engaging in ‘censorship by proxy’. As per Article 12 of the Constitution, whenever a government body performs a ‘public function’, it must comply with all the enshrined fundamental rights.

Any individual has the right to file a writ petition against the state for violation of a fundamental right, including the right to free speech.

However, judicial precedent on the horizontal application of fundamental rights, which might enable an individual to enforce a similar claim against a private actor has not yet been cemented in Indian constitutional jurisprudence.

This means that any individual whose content has been wrongfully removed by the platform may have no recourse in law — either against the state or against the platform.

Algorithmic governmentality

Using automated technologies comes with its own set of technical challenges even though they enable the monitoring of greater swathes of content. The main challenge to automated filtering is the incomplete or inaccurate training data as labelled data sets are expensive to curate and difficult to acquire, particularly for smaller players.

Further, an algorithmically driven solution is an amorphous process.

Through it is hidden layers and without clear oversight and accountability mechanisms, the machine generates an output, which corresponds to assessing the risk value of certain forms of speech, thereby reducing it to quantifiable values — sacrificing inherent facets of dignity such as the speaker’s unique singularities, personal psychological motivations and intentions.

Possible policy prescriptions

The first step towards framing an adequate policy response would be to segregate the content needing moderation based on the reason for them being problematic.

Detecting and removing information that is false might require the crafting of mechanisms that are different from those intended to tackle content that is true but unlawful, such as child pornography.

Any policy prescription needs to be adequately piloted and tested before implementation. It is also likely that the best placed prescription might be a hybrid amalgamation of the methods outlined below.

Second, it is imperative that the nature of intermediaries to which a policy applies are clearly delineated. For example, Whatsapp, which offers end-to-end encrypted services would not be able to filter content in the same way internet platforms like Twitter can.

The first option going forward is user-filtering, which as per a recent paper written by Ivar Hartmann, is a decentralised process, through which the users of an online platform collectively endeavour to regulate the flow of information.

Users collectively agree on a set of standards and general guidelines for filtering. This method combined with an oversight and grievance redressal mechanism to address any potential violation may be a plausible one.

The second model is enhancing the present model of self-regulation. Ghonim and Rashbass recommend that the platform must publish all data related to public posts and the processes followed in a certain post attaining ‘viral’ or ‘trending’ status or conversely, being removed.

This, combined with Application Programme Interfaces (APIs) or ‘Public Interest Algorithms’, which enables the user to keep track of the data-driven process that results in them being exposed to a certain post, might be workable if effective pilots for scaling are devised.

The final model that operates outside the confines of technology are community driven social mechanisms. An example of this is Telengana Police Officer Remi Rajeswari’s efforts to combat fake news in rural areas by using Janapedam — an ancient form of story-telling — to raise awareness about these issues.

Given the complex nature of the legal, social and political questions involved here, the quest for a ‘silver-bullet’ might be counter-productive.

Instead, it is essential for us to take a step back, frame the right questions to understand the intricacies in the problems involved and then, through a mix of empirical and legal analysis, calibrate a set of policy interventions that may work for India today.