The Hidden Opportunity to Regulate Targeted Political Advertising in India

Political parties were big spenders on online political advertising in the recent Indian General Elections, with a report from April 2024 detailing that the Bharatiya Janata Party spent Indian Rupees 23.04 crores, with the Indian National Congress being the second highest spender at Rupees 16.07 crores. But, while there is scrutiny on how much political parties might be spending on advertising, there is less scrutiny regarding how such targeted political advertising is delivered, including the surveillance required to make it effective. However, the Supreme Court of India’s recent verdict in Association for Democratic Reforms v. Union of India has created some inroads for these conversations to take place in India, hopefully leading to the regulation of targeted political advertising.

How does it work?

Before getting into the Supreme Court’s verdict, it may be useful to recap how such advertising works. Targeted advertising, or microtargeting, involves selecting a pre-defined group of users based on certain attributes (such as gender, age, or location) and delivering the desired advertisements to these users through a platform that acts as an intermediary between the user and advertiser (in this case, the political party). The algorithmic process through which these ads are delivered is largely a black box, with historical behavioral data aggregated across multiple thousands of users to determine when and to whom these ads are delivered - studies have shown the potential for the impairment of consumer autonomy through targeted advertising.

While the algorithms may lack transparency, what is known is that vast amounts of data are required to make the delivery of these ads effective, which means constant tracking of users' behavior online. Even contextual advertising which is touted as less invasive, is being implemented using the latest AI technologies, device fingerprinting, session data, and location-based targeting (among several other data points), creating the same potential for manipulation as behavioral advertising.

What is particularly worrisome about political advertising through microtargeting is the potential for influencing voters’ choices by delivering misinformation and fake news that is designed to appeal to specific users or communities and inundating them with information that may distract them from more pressing concerns.

The privacy of political affiliation

In India, these conversations are fairly nascent. The government introduced a standalone data protection legislation only in 2023 after the Supreme Court of India held that citizens have a fundamental right to privacy under Article 21 of the Indian Constitution in 2017.

In a recent development in February 2024, a challenge to the ‘Electoral Bonds Scheme’ came up before the Supreme Court. The Scheme created a mechanism by which political parties could receive unlimited funding from individuals and corporations under complete anonymity, even exempted from the scope of India’s right-to-information laws. In a unanimous verdict, the Scheme was struck down by the Supreme Court as violative of the Indian Constitution.

Interestingly, while striking down the Scheme, the Court considered whether the right to a donor’s privacy could be grounds for a voter to be denied information about the source of contributions to a political party. In discussing this issue, the judges considered whether the fundamental right to privacy under the Indian Constitution (as mentioned above) would encompass the privacy of one’s political affiliation. They ultimately concluded that such privacy would need to exist to protect those making bona fide donations as an expression of support for the political party in question. The rationale was that all donors are not wealthy individuals or corporations, and many are average individuals who are exercising their democratic right to show support for their political party of choice.

The Court also observed that privacy of political affiliation is important since surveillance can be used to disenfranchise voters. For instance, this can be done through analysis of voter databases to identify voting patterns and target voters based on these patterns, collecting data from online behavior and purchase histories, and tracking news consumed to understand the political leanings of these voters. What the Supreme Court does here is acknowledge, for the first time in India, the possibilities of algorithmic surveillance in the context of elections (although it is to be noted that there have been earlier references to online tracking and data collection in the judgment upholding the fundamental right to privacy).

The significance of this verdict

By expressly reading in a right to the privacy of one’s political affiliation, the Supreme Court creates a whole host of possibilities for the future of regulating targeted political advertising. If the online surveillance of voters to conduct analytics has the potential for disenfranchisement and the exercise of undue influence on the voting process, then surely targeted political advertising falls within this ambit.

A report submitted to The Guardian reveals that in the 2024 General Election, Meta approved a series of AI-generated political ads through Facebook and Instagram, which spread disinformation and incited religious violence. The report also reveals that Meta failed to block advertisements targeted to users in highly contentious districts during the ‘silence period’, (the last 48 hours before polling when advertising is prohibited). Researchers found that both major political parties (i.e., the BJP and Congress) failed to comply with prohibitions on ads in the ‘silence period’ and utilized ‘substantial sums’ to sway voters with ‘hyper-targeted’ ads.

In Shivam Shankar Singh’s book, How To Win An Indian Election: What Political Parties Don’t Want You To Know (excerpted here), the author details how huge public datasets, such as electoral rolls, can reveal names, ages, and gender, and can then be matched with parameters such as religion and caste with a mix of technological tools and manual labor. Land records, Below Poverty Line lists, and electricity bills, for example, can provide indicators of socioeconomic status. Once this large dataset is created with relevant indicators, political parties can obtain phone numbers from data brokers, after which specific groups can be created to target certain types of messages on WhatsApp, through messages, and through platforms like Facebook. This, the author says, “would allow for targeted messaging and micro-targeting in a way that hasn’t happened in any other country in the world.”

Clearly, all these practices interfere with the autonomy of the voter to influence their votes, which is an ‘illegitimate practice’ which can ‘be used to invalidate the foundation of the electoral system’, as per the Supreme Court. This logically leads to the conclusion that the very existence of targeted political advertising without the necessary level of oversight violates one’s privacy of political affiliation in the manner conceived in this case.

Going a step further, I argue that the broader issue beyond the possibility of disinformation, fake news, or other such adverse effects is the ethics of targeted advertising itself. What matters is not so much the consequences as the allowance of a practice that leads to the surveillance of millions of users through tracking. With the Supreme Court having acknowledged that voters have a right to have their political affiliation safeguarded (which importantly includes drawing inferences as well), this is a golden opportunity to challenge the current status quo. We need a larger conversation on whether Indian voters should be subject to targeted political advertisements that rely upon personal information that the Supreme Court says should be protected.

What India can build on: An autonomy-based approach

While the matter before the Supreme Court was restricted to deciding on the constitutionality of the Electoral Bonds Scheme, the recognition of online tracking as adversely affecting the privacy of political affiliation could have gone further - for example, by recommending that rules be framed to regulate the discriminate use of data for political advertising (similar to how the Supreme Court recommended that a data protection law be enacted to give effect to the fundamental right to privacy).

This is not to say that the Election Commission of India hasn’t taken steps to regulate political ads at all. In a press note issued in April 2024, the Election Commission reminded political parties of guidelines released in 2013 specifically pertaining to political ads on social media. These guidelines require candidates to provide details of their official social media accounts to the Commission and mention that all political ads on any platform need to be pre-certified by the Commission and are subject to the same Model Code of Conduct as other communications during campaigning. Political parties are also required to maintain and submit statements of expenditure on such ads.

However, technology has moved much farther since 2013, and this does not go far enough to address the root of targeting techniques. India requires a framework to prevent such data from being indiscriminately used in the first place, as well as restrictions on the use cases for targeted advertising when such advertising could subvert the democratic process. The Supreme Court’s judgment more than provides the basis for these long-needed regulations to be introduced and for a legal challenge to be brought to direct the government to introduce these regulations.

The EU has recently introduced rules on transparency and targeting of political advertising. These rules require political advertisements to contain a transparency label with an easily retrievable transparency notice, which must clearly identify political advertisements as such and provide some key information about them, including their sponsor, the election or referendum to which they are linked, the amounts paid, and any use of targeting techniques. Targeted political advertising will only be permitted with the explicit consent of users for political advertising, and sensitive categories of data for this purpose (such as racial or ethnic origin) cannot be used for profiling.

These are elements that can improve the Election Commission's erstwhile rules on the subject. However, we can also go further - by ensuring that our data protection legislation covers profiling and the usage of behavioral data (I have previously argued why India’s Data Protection Act lacks in this regard), imposing spending limits on targeted advertising, or yet, even prohibiting targeted advertising for certain purposes which have democratic importance, altogether. All these measures would be entirely compatible with the Supreme Court’s interpretation of the privacy of political affiliation.

But, what is even more necessary is to move beyond looking at this issue in a consequentialist manner - what matters is the grander principle of voter autonomy. In a country like India, with extremely poor voters, a significant digital divide, and very low rates of digital literacy, the population is far more susceptible to being taken in by fake news and disinformation that sparks violence. Therefore, the privacy of political affiliation in the Indian context goes beyond seeking consent and transparency labels (which most internet users in the country wouldn’t even be cognisant of) - it is essential for the protection of the most vulnerable in the world’s largest democracy.