Why EU must limit political micro-targeting

The advertising machines of social networks, video portals and search engines make it relatively easy and cheap to reach masses of voters quickly.

While this has helped increase the scale of voter outreach, the unprecedented masses of ads make public-interest scrutiny hard for journalists, researchers and regulators.

Political advertisers can also use the platforms' algorithmic ad delivery to find specific audiences and hammer home certain political messages over and over again.

This micro-targeting helps campaigns reach voters but, again, comes with risks: there is a danger of paying to amplify existing polarised debates and voter segmentation. These risks need to be urgently addressed by putting in place safeguards for micro-targeting.

Limiting micro-targeting is not a new idea, but is gaining traction in the EU, as it is part of the deliberations surrounding the European Democracy Action Plan (EDAP).

The EDAP deals with questions of media pluralism and political advertising, and one key component could be limiting micro-targeting for political ads.

'Micro' does not mean tiny

Micro-targeting is not necessarily equivalent to targeting tiny audiences, it means targeting specific, homogeneous audiences.

This is possible on social media due to the vast amounts of personal behavioural data that platforms have on users. Such personal data is used to infer people's preferences, dislikes and fears, and to group them into advertising profiles.

Advertisers can then opt to have platform algorithms deliver ads to those people most likely to engage with their political messages.

It remains hard for researchers to study micro-targeting and its effects, in part due to a lack of access to companies' data.

Nonetheless, campaigns in the US and across the EU as well as other parts of the world have shown how targeting might skew political debates and agendas: if a campaign can figure out exactly what certain groups of voters want to hear, they can pay to repeatedly send that message to them.

In contrast to campaign posters on the street or ads on TV, other voters have no way of knowing what paid political messaging their fellow citizens are seeing. People become siphoned off from each other and societal divides might harden.

To be clear, limiting micro-targeted advertising is not a panacea for all ills afflicting online political campaigning. A whole range of measures is necessary to address shady ad practices – starting with improved access for independent researchers and mandatory transparency measures for platforms and advertisers.

Yet, finding ways to put a brake on micro-targeting has several advantages: it could reduce the amount of personal data that is used for paid political messaging, because platforms would have less incentives to collect personal data in the first place.

It could allow counter-speech and more public-interest scrutiny, because ads would have to be targeted to larger, more heterogeneous groups. And ultimately, the hope is that this could lower the risk of political ads further polarising the electorate.

There are different options to limit micro-targeting that should be discussed during the development of the EDAP as well as the upcoming Digital Services Act (DSA), which deals with platform transparency and accountability.

Civil society experts and academics have already put forth ideas and some of the European Parliament's committee draft reports on the DSA make specific proposals as well.

It seems promising to put limits into place as to what and how much data can be used for creating users' ad profiles. That way, it is harder to target homogenous groups.

It could be debated whether a minimum size for political target audiences could be implemented, which would further enhance options for counter-speech.

There could also be an opt-in for users to receive targeted advertising. Expanded disclosures of ad targeting and delivery criteria would benefit users and allow better public-interest scrutiny.

Platforms already do some of these things: for example, Facebook offers a tool to review what ad interests users have and Google allows only basic demographic data to target political ads.

Such laudable voluntarily efforts should be improved upon, made mandatory and implemented industry-wide.

The EU should now seize the opportunity to tackle risks associated with paid political messaging online and take the lead on this to set a standard for globally operating ad delivery platforms.