skip to Main Content

Today, Vice President and Commissioner for Values and Transparency Jourova presented the European Commission’s proposal for a regulation on the transparency and targeting of political advertising. Online political advertising is a relatively new phenomenon that has not been thoroughly regulated – a regulatory gap which threatens electoral integrity through voter manipulation, intransparent campaigning and disinformation campaigns, as shown by the Cambridge Analytica scandal and recent whistle-blowers’ testimony. This regulation acts as lex specialis to the Digital Services Act to focus specifically on political advertising online, overcoming the failures of the self-regulatory Code of Practice 1.0. In the last years, we have been following this file closely, pushing for a regulation with a policy dialogue regarding online political ads, a statement on political ads and a lengthy input paper to the European Democracy Action Plan. Here we share our initial reaction to the proposed legislation.

The proposal by the European Commission is a stepping stone towards ensuring fair and transparency online campaigning. The whole-of-value-chain approach that translates on imposing obligations on the whole value chain of political advertising and not only on platforms will make the regulation more effective. Also, leaving most technical details in annexes subject to review in delegated acts makes it future-proof. However, minor additions to the proposal would lead to much greater positive impact in terms of electoral integrity, trust in democracy and protection of fundamental rights. Therefore, we propose to improve this regulation by including:

  • A mandate for universal, real-time, comprehensive ad libraries;
  • A ban to the use of inferred data for political advertising; and
  • Further clarifying the definition of ads liable to influence a political process.

The success of the regulation will depend to a great extent on:

  • The resources devoted to the implementation of the regulation by platforms and very large online platforms (VLOPs) in particular;
  • The introduction of proportionate, dissuasive, and effective sanctions in all Member States; and
  • The willingness of the co-legislators to be more ambitious than the current proposal.

In the absence of a common definition of political advertising or political advertisement, it is necessary to provide a common definition (Article 2.2) to ensure the effectiveness of this initiative. This will empower national electoral bodies to engage with platforms much more effectively, while still providing space for national variations in national electoral law. The mostly actor-based definition is greatly welcomed, as it covers both political parties and those paid by political parties (for instance influencers) to campaign.

While the regulation is clear on ads by political actors or those advertising on their behalf, we enter murky waters when it comes to issue-ads by CSOs or corporations. The Commission proposes that ads are political if they are “liable to influence the outcome of an election or referendum, a legislative or regulatory process or voting behaviour”. While this detailed list of criteria is a helpful step forward, the definition stands and falls with the mechanisms for identifying an ad as political. If in practice this definition means that platforms may employ a list of ‘political issues’ linked to elections or legislative processes that are automatically marked as political ads, this could unnecessarily restrict fundraising or mobilisation activities of NGOs working on LGBTQ+, migration or climate issues – topics that have typically been controversial around elections. Moreover, the question remains whether such a definition will capture the disinformation and foreign interference campaigns the Commission is targeting.

Further clarification on the interpretation of when ads are liable to influence elections, referenda, legislative or regulatory processes or voting behaviour is needed to prevent arbitrariness in the enforcement and undue restrictions on public interest campaigners. Detail is lacking on the resources dedicated by platforms – particularly (VLOPs) – to classifying political and commercial ads. This will require human supervision and rapid response to flaggers.

We welcome the introduction of viewer-level transparency obligations beyond the obligations foreseen in the DSA (Articles 13 and 30) under articles 6 and 7. Core information will be provided in the ads and additional information will be available over transparency notices in the ads We believe that this will greatly contribute towards solving a number of issues. Most importantly, citizens will be able to link the ads they see to the actor on whose behalf it is published and its political and sponsored  nature, see how much they spent on an ad, why they are targeted with an ad and what  data source was used for this targeting, with a link to add libraries for Very Large Online  Platforms. The data types provided over transparency notices are subject to be expanded  in a delegate act, which might be necessary to respond to fast technological  developments.

First, the proposal does not include an obligation for real-time transparency, which is essential in a fast-paced electoral environment. Secondly, the proposal misses a  chance to provide meaningful transparency not on individual ads, but on advertising  campaigns as a whole, accessible to watchdog CSOs and media. The most egregious  violations of electoral rules and ethics in campaigning, as well as disinformation  campaigns, involve multiple ads, not a single one. Journalists and CSOs have used  existing ad libraries (despite their flaws) for investigations on political campaigning  that have exposed harmful practices and malign influence operations. Public scrutiny  is key to holding political parties to account at a critical time for democracies, which  requires easy, real-time, public access to ad repositories. The lack of campaign-level  transparency obligations comes despite the fact that the regulation acknowledges in  recital (44) that aggregate data on expenditure in campaigns can contribute to political  debate.

Therefore, as called for by ERGA and multiple consulted stakeholders, we call for a  mandate for real-time, comprehensive ad libraries for political ads that would contain  all the information listed in Article 7.1, Annex I and Annex II for each and all political  ads. Data must be provided in detailed numbers avoiding the use of ranges.

  • This would be much more cost effective than the complex system that results from the combination of Article 7.6, Article 8 and Article 11 for the provision  of aggregated data. Technical solutions involving a small one-off cost could be  developed to automatically transfer all information shared with the viewer as  per Article 7.1, Annex I, Annex II to an ad library. Platforms which are not VLOPs  and do not have an obligation to have an ad library as per the DSA, could host  information on the ads they have published in libraries of their own, ad  libraries of VLOPs or an ad library provided by the European Commission. This  mandate would ensure that political actors are held accountable in real time if they exceed expenditure limits of if the contents of their ads is misleading  or contradictory.

The regulation builds upon the GDPR and strengthens its enforcement and  compliance. This will be achieved by introducing requirements for targeting,  essentially mandating greater transparency in the processing of personal data, and  reiterating the ban of the use of special categories of personal data as defined in the  GDPR (Art 9). Following the GDPR logic, this ban does not apply when the user data  “has given explicit consent to the processing of those personal data for one or more specified purposes”.

Enhanced transparency requirements work well for targeting techniques, but not so much for ‘amplification techniques’ (Article 2.8).

  • Targeting (addressing a political advertisement only to a specific group of persons) can be made transparent in a meaningful way. The criteria used by publishers at the sponsor’s direct request to target viewers are specific, limited, traceable, and deliberately chosen, so, upon disclosure, individuals can understand these criteria.
  • On the contrary, the functioning of amplification techniques, this is, the algorithms used by platforms to optimise the circulation, reach and/or visibility of ads, are inherently immune to meaningful transparency. Platforms’ ad delivery algorithms, which are more powerful and potential pernicious than targeting techniques to determine who sees which specific ad, involve profiling users processing thousands of apparently irrelevant parameters to deliver ads to them in an optimised, profit-driven manner. It cannot be expected that viewers of an ad will meaningfully understand why and how the platforms’ algorithms delivered ads to them even if the algorithmic process would be made transparent.

Sufficiently informed and explicit consent to the processing of data – and inferred data in particular – is often missing due to the generalised use of dark patterns. Furthermore, there is evidence of the processing of inferred data aimed at exploiting the vulnerabilities of citizens.

To address this, we propose a ban on the use of inferred data for all political  advertising, as called for by the European Data Protection Board. There is legal basis  for prohibiting the use of data for this purpose. Given the particularly pernicious uses  of inferred data by platforms to deliver political content sponsored or not – to users  as most recently revealed by Facebook whistle-blower France Haugen, a ban on the  use of inferred data in political advertising must be adopted

The requirements of “explicit consent” for the processing of special category data for  political advertising purposes could be open to interpretation by different political and  economic actors. Further clarification on this issue would be needed to enhance legal  certainty and protection of fundamental rights.

We welcome the enforcement regime in the proposal. The enforcement regime of  the GDPR already in place and the upcoming enforcement regime for online  advertising transparency foreseen in the DSA are best suited to enforce the obligations  included in this regulation. We also welcome the provision of clear guidelines to  Member States on the establishment of sanctions. Since the failure of a single  Member State to introduce dissuasive sanctions would emperil the success of the  Regulation, these ambitious guidelines are a step in the right direction. Sufficient  resources must be allocated to national and European organisations enforcing data protection and online advertising transparency, accounting for the increased  workload that this regulation will imply. It must be noted that data protection bodies  in particular are reported to be understaffed and underfunded.

The success of this legislation will depend on Member States adoption of sanctions in  line with the guidelines provided in Article 16 (effective, proportionate and dissuasive  in each individual case). As online advertising services are provided on a cross-border  basis, it would only take a single Member State not to adopt adequate sanctions to  jeopardise the success of the regulation.

More news

Back To Top