Online political ads - a study of inequality in transparency standards

Over the past years, there has been much attention paid to political advertising transparency on online platforms. Growing internet penetration and the rising popularity of social media have made social media platforms a key battlefield for political actors in the fight for votes, where political ads have proved themselves to be a popular weapon. This means online transparency towards voters has become more imperative than ever.

And yet, social media platforms’ regulation of political ads remains fragmented. Each platform operates fluctuating and often widely differing transparency standards for different countries. These differences operate at multiple levels. For a start, each platform defines political content and ads differently, with Google going as far as operating a country-specific definition of electoral ads. In Facebook’s case, there is also the regulation of “social issues” - issues that are not directly related to political campaigning but are still politically sensitive (e.g. immigration). Social issues vary from one country to the next, with little clarity as to the criteria applied by Facebook. Unless an ad is regulated as “political” or a “social issue”, next to no information is provided to users as to the spend, reach, targeted audience and advertiser of a specific ad. In many countries, this regulation doesn’t take place, leaving both individual users and civil society in the dark.

The combined effect of these inconsistent policies is to create a two-tiered system within each platform, with a wide regulatory gap separating the transparency haves and have-nots. We highlighted these problems in 2019. While some progress has been made since, we are saddened to observe the transparency divide persists. And as our research shows, it has consequences for democracy.

Background

PI commissioned in-depth studies on political ads transparency in two if the world’s largest democracies: Brazil and Indonesia. Our partners InternetLab and ELSAM looked into Facebook, Google and Twitter’s definitions of political and social issue ads, how these platforms were deploying transparency tools for political advertising in these countries (if at all), the adverse impacts of non-existent or limited transparency, and what could be done to improve the standards.

We then carried out a comparative analysis of the electoral regulatory framework and online platform policy development in these two countries. To inform our research, we relied on publicly available information regarding the full range of transparency tools that the platforms make available to different countries, company policies, and reporting. We also asked Facebook and Google for clarification and have included wherever relevant their responses.

Findings

Below is a brief summary of our research findings.

Heightened transparency standards are a privilege of the few

The US election was a clear illustration of the transparency divide between haves and have-nots, as well as its ramifications. Google applied no less than two “sensitive event” policies resulting in temporary bans of political ads, and Facebook carried out at least five significant changes covering voter and census participation suppression, additional transparency checks for candidate spend, increased verification requirements for branded content, and user controls to see fewer political ads. In Facebook’s case, some of these changes were part of a specific - and seemingly globally unique - plan to “protect the US election”. The extent of these changes and the speed with which they were implemented is, to PI’s knowledge, unrivalled in any other country.

The availability of transparency tools does not equate with their use

Facebook makes its transparency tools available globally. In some countries, such as the US and the EU, these tools are compulsory for advertisers to use. In others, they remain optional. With little incentive to comply with these optional tools, advertisers don’t use them.

There is little clarity about the criteria used by any platform to regulate political ads in one given country

PI could not find any publicly available information detailing the internal processes by which social media platforms decide to regulate political ads in any given country. The difference in treatment can be stark. In Peru’s 2020 congressional elections, political advertisers on Facebook were allowed to freely targets ads to users located in Peru without checks. However, the same advertisers were prevented from targeting ads to US-based users on account of having failed to comply with the mandatory transparency tools enforced in the US.

In the absence of ad repositories, harmful content can go undetected and unscrutinised

Across Facebook and Google, the regulation of political ads necessarily results in those ads being automatically recorded in ad repositories (Facebook’s Ad Library, Google’s Transparency Report), with detailed information about the ad. However, in the absence of regulation, ads seeking to exploit controversial political issues can fall through the cracks. This in turn makes it harder, if not impossible, for civil society to monitor any potential harms on democracy.

Apps such as WhatsApp and Instagram are becoming relevant political advertising platforms

Though lesser known as political campaigning arenas, the use of political ads in WhatsApp and Instagram has significantly increased. It is possible for Instagram ads to be included in the Ad Library - but only if they are shown in countries which benefit from political ads regulations.

Recommendations

Based on our findings above, we articulate the following recommendations:

• Online platforms must ensure that they provide heightened, compulsory transparency standards globally.
• Social media companies must ensure that they provide similar transparency standards across all platforms which are used to publish political content.
• Online platforms must adapt their transparency policies to capture all types of political advertising, and take steps to ensure that these policies are enforced.
• Online platforms must make public the criteria they use to deploy heightened transparency requirements.
• Online platforms must commit to higher standards than the basic requirements laid down by electoral law.
• Online platforms must ensure that they engage in open dialogue with electoral and data protection regulators. Cooperation between online platforms and regulators is key to ensure that the extent of online political advertising and related expenditure by political actors is known and scrutinised.

This research is a part of PI’s work on Defending Democracy and Dissent, which aims to investigate the role technology plays in facilitating and/or hindering everyone’s participation in civic society.