Should online political advertising be regulated?

by P.J. George — last modified Nov 13, 2019 03:12 PM
Micro-targeting could have potentially damaging results in the context of political advertising.

The article by P.J. George was published in the Hindu on November 8, 2019. Pranesh Prakash was interviewed.


On October 31, Twitter announced that it will no longer carry political advertisements as the power of Internet advertising “brings significant risks to politics, where it can be used to influence votes”. On the other hand, Facebook has said it will not fact-check political advertisements as it does not want to stifle free speech. In a conversation moderated by P.J. George, Pranesh Prakash (board member, The Centre for Internet and Society) and Kiran Chandra (General Secretary, Free Software Movement of India) discuss how platforms and constitutional authorities can deal with the challenges posed by online political advertising to democracies. Edited excerpts:

We have always had political advertising. What is it that makes online political advertisements different or maybe even problematic?

Pranesh Prakash: There are two things that make online political advertising different. One is targeting. Online advertising allows, especially on social networks, for a kind of targeting that wasn’t possible at the same level before. Earlier, if you wanted to target a particular segment of people for your political messaging, you could find out what kind of magazines they subscribe to and put fliers in those magazines. But you couldn’t engage in personalised targeting based on multiple attributes that is possible through platforms like Facebook and Twitter. The second is the invisibility of this kind of advertising. If there’s a billboard in the real world, everyone gets to see it. However, if there’s targeted advertising on a social media platform, not everyone gets to know of it.

Kiran Chandra: App-based organisations have designed advertisement models to specifically allow targeting. Facebook, for instance, allowed you to choose a person from a particular caste and also from a particular class in the same caste. If somebody wants to look at an advertisement for an Audi, they can go to one class of newspapers or look at billboards in some localities; the very existence of the product is not opaque to society. But targeted advertising makes it possible for two people connected to the Internet from the same source, using the same equipment, studying in the same school or college, working in the same workplace, and living in the same habitat to get two different advertisements. And micro-targeting has got potentially damaging results in the context of political advertising, particularly for elections. These platforms make it possible to go from manufacturing consent to manipulating consent. A person is continuously fed with information to vote for a particular party.

Twitter said it will no longer carry political advertisements, considering the repercussions seen in the U.S. in the past elections. On the contrary, Facebook says political advertisements are necessary and that people should see if their politicians are lying. How culpable is a platform in the case of a problematic online political advertisement?

KC: Platforms, particularly Facebook, have been washing their hands of the issue saying they are only intermediaries providing space; that the content is being generated by the people to be consumed by the people, and they have no role to play. But this is false. If you look at the complete business model of Facebook, Google, or any of the platforms, they clearly provide micro-targeting, or allow people to be manipulated for a particular purpose. So, these platforms can’t just wash their hands of the issue. In the Maharashtra election, you saw a lot of advertisements coming out which are untraceable. How can this happen without the platform itself allowing for such a possibility? The Election Commission (EC) needs to step in on all these issues. These corporations need to be very transparent in the context of elections. They need to bring out all the ways in which advertisements are displayed and also the money associated with it.

When somebody publishes it [an ad] on a Facebook wall, it is as good as publishing it in a newspaper. So, all the legislation that apply now for reasonable restrictions and freedom of speech and the freedom of press also apply to these platforms. These platforms are culpable when the very intent of their business model allows such subversion of the democratic process. They need to be brought in line to ensure that Indian democracy is safe.

PP: I completely disagree with Kiran on a number of points. For instance, those who are running a platform shouldn’t automatically be liable for what people are seeing on those platforms. The people who are actually saying things should be liable, not necessarily those who are carrying it without knowing what they’re carrying most of the time. Kiran also mentioned manipulation. The job of all advertising is to manipulate. The job of newspapers is to manipulate public opinion. And there’s always money associated with this. Newspapers carry advertisements as well. You don’t necessarily know who has paid for each ad in the newspaper. What online platforms are able to provide is actually greater transparency in this regard, at least based on what Facebook is attempting to do with its ad library. Calling this manipulation doesn’t quite work. Because then you have to specify why certain categories of things you think of as manipulating, while other categories you think of as influencing.

Second, as far as I know, Facebook does not ask for your caste. Nor does it actually allow advertisers to use caste as a category for advertising. To address the larger question of whether to carry political advertisements or not, I don’t think there are simple answers. For instance, in different jurisdictions there are different rules as to whether different kinds of media are allowed to carry political advertisements or not. In the U.S., all broadcasters are required by law not to censor on the basis of the content of political advertising. Which means that broadcasters in the U.S. cannot say to a candidate, ‘this advertisement that you’ve sent to us contains a lie and we’re not going to associate ourselves with the lie and we’re not going to carry it’. Now, when a platform like Facebook says that it will voluntarily adopt a similar standard as applies to broadcast organisations by law, all hell breaks loose. And again, there might be good reasons for it. But to say that political advertising should not contain lies, and hence should be censored, is not a viable opinion across the board.

KC: I would like to clarify one thing here. There is a clear distinction between Facebook asking your caste and Facebook allowing you to micro-target people based on their caste and class. In 2016, I created an advertisement with a tag called Brahmin bags and it allowed inclusion and exclusion based on caste and economic status. And now, after this had been made an issue for the last three years, Facebook says that advertisers can select topics that are specific to a particular caste. For instance, Dalit topics, Iyengar topics, etc. So Facebook, in its design, allows such kind of sensitivities to be used for micro-targeting. And one should not confuse general advertising with political advertising. If the advertisement is just about manipulating for buying a particular product, that has something to do with the business houses; even if one agrees with it or not. But when you speak about political advertising, when people come to participate and engage in a democratic process, the EC and The Representation of the People Act (RPA) mandate that people should be allowed to take a very clear stand, to look at what has happened in the last five years, and decide how to vote, freely and fairly. That is why the RPA clearly lists a certain set of things for free and fair elections, where even the use of money and manipulation should not be allowed to happen. Yes, the U.S. has a different context. American democracy is different from Indian democracy. We have got our own statute. This methodology in which these platforms have got their business models and are engaging deeply in subverting the Indian democratic process is a serious cause of concern. The EC should come up with new methodologies, if the existing ones are not sufficient.

Can you elaborate on how the EC can play a role in this?

KC: We brought these issues to the notice of the EC prior to the 2019 general election. The EC said it does not have enough manpower to deal with this situation for now. The EC does not have power over the police or the administration; but once the elections are on, it has the capability to take in different departments and ensure that such subversion of the democratic process does not happen. A fundamental problem with the EC’s method is that it said it was in discussion with the digital platforms to make more people vote in the election. And that itself is problematic. How is it going to be done? The EC should make public the way in which this advertising is being conducted, the money associated with it, and the people who are being reached with it. For instance, if we look at TV channels for ads during primetime, there is a mechanism, like TRP ratings, which allows them to understand and evaluate the target sections. If you look at the Maharashtra election, the advertiser itself is not known. Have people been sent communal messages? Have people been targeted based on caste, which can disqualify the contestant? The EC should reach out to the Government of India and look at the departments that are capable of handling this. If they don’t exist, it should start creating infrastructure that will be able to look into all these aspects. Also, concrete guidelines should be given to these digital platforms. And whatever comes in contradiction, or comes in the way of implementing the RPA, the EC should stop the platforms from doing it.

PP: For me, it’s not clear to what extent I would draw a distinction between advertising and other things which the EC has not been able to curtail, such as paid news and political ownership of media, which allow for very skewed viewpoints to be expressed. But insofar as what can be done about online platforms — and again, only online platforms which deal in advertising — the biggest source of online political messaging in India is WhatsApp. So, excluding the elephant in the room from this discussion, what the EC could do is bring the largest platforms together to get transparency commitments from them. Then this information needs to be made publicly available, so that the invisibility which happens with targeting gets countered. The second thing... Given that elections are geographical in nature in India, if you want to engage in advertising, you have to do it on the basis of geography, not on the basis of specific kinds of attributes of a person. And let’s also be aware that most of these attributes or guesses about people that these platforms are making are based on what people post on social media platforms, what they click. So, the one thing that can be done on a global level is transparency and restrictions on various targeting but anything else such as limitations on, say, lying in political advertising, I don't think that can or should be sold on a global level. It’s dependent far too much on each country and their models and how they interpret freedom of expression.