The EU is going too far with political advertising!
Max van Drunen, Natali Helberger, Wolfgang Schulz, and Claes de Vreese
The EU is set to complement the DSA with a new regulation on (targeted) political advertisements. In this piece we highlight how the regulation’s definition of political advertisements and enforcement mechanisms threaten freedom of expressio
Right now, the exclamation above is just a title. Depending on the outcome of the final negotiations on the EU’s proposal for a Political Advertising Regulation over the following months, this title could soon be considered a political ad, and subject to regulatory scrutiny. This is due to two issues with the regulation’s definition of political advertisements. The regulation takes a very broad approach to determining what is political, and includes any message that is liable to influence voting behaviour. Moreover, the regulation takes a counterintuitive approach to defining advertising: its rules on targeting potentially also capture political speech that is not part of a commercial advertising service.
The result could be a regime that encourages censorship, chills certain forms of political speech, and burdens political speakers with far-reaching limitations on distributing their messages. Though the details are still under negotiation, the regulation’s general approach is becoming clear. The regulation will require advertisers to declare whether their ad is political, and provide extensive information about (among other things) their identity and the way the ad was funded. Publishers (like platforms, newspapers, and TV stations) must relay this information to the public and label the ad as political. If it is self-evident to publishers that an advertiser has failed to correctly declare a political ad or provided wrong or incomplete information – and the advertiser fails to correct it – the ad may be removed. Users may also report incorrectly labelled political ads, and in the month before an election, publishers must address these notifications within 48 hours. Targeting young people or making use of sensitive data without consent will likely be prohibited. The EP has proposed further restrictions on political microtargeting, including a ban on any data that is not explicitly provided by the user with specific consent for its use for the purpose of political advertising.
The intentions are good – strengthening the democratic process by ensuring fairness and transparency in political advertising. But the key question is to which and whose messages these rules will be applied.
What is political advertising?
The regulation would not only apply to messages from politicians and political campaigns, but also to so-called issue ads – messages from any actor that are “liable to influence the outcome of an election … regulatory process or voting behaviour”, in the words of the original proposal.
This is an exceptionally broad definition. The European Parliament (EP) and Council have added language to the recitals that would broaden it even further. Both point out that commercial advertising can use social issues to promote products and brands, and that it too falls under the Regulation, if those commercial ads are liable to influence voting behaviour. The EP even goes so far as to argue that any message that is liable to influence “public opinion on societal or controversial issues” should be considered political advertising. Note also that the definition of political advertisements itself does not require that someone pays for the message. The European Parliament closed this loophole in the Commission’s proposal by clarifying that the entire regulation only applies to economic services. However, the Council explicitly states the restrictions on targeted political speech should apply no matter whether a service is involved or not. As over 30 NGOs emphasised in an open letter to the Council, this creates serious challenges for civic speech.
It is easy to see how the political advertising regulation’s definition can capture ads that are not traditionally seen as political. Commercial ads that try to capitalise on societal issues like Kendall Jenner’s Pepsi ad may have many problems, but none that need to be addressed with political advertising law. Ads from an NGO that spread the word about climate change, or fundraise to counter new legislation that threatens media freedom, would be covered as well if they are liable to influence regulatory processes or voting behaviour. This opinion piece will only escape regulatory scrutiny if it falls under the exception that the Commission included in the recitals (and the Council and EP would make binding law) for views expressed under the editorial responsibility of a media service provider without remuneration. The breadth of the regulation is perhaps best signified by the fact the EP found it necessary to clarify that catering, cleaning, and maintenance are not political advertising services.
How will the political advertising regulation affect non-politicians?
All of this is bad news for civil society organisations or individuals that participate in the public debate. If NGOs take the safe approach and declare their message a political ad, it will be labelled as such for users. The NGOs will have to comply with the regulation’s extensive transparency obligations, and their ability to target audiences will be restricted. In particular the restrictions on targeting young people and targeting people based on their political opinion may severely limit NGOs’ ability to reach their constituents. The EP’s suggestion to limit targeting to data users have explicitly provided for the purpose of political advertising limits NGOs’ ability to reach their audience even further.
Advertisers that do not declare their ad as political, risk violating the regulation. This is where the vagueness and broadness of the definition of political advertising really begin to bite. The regulation creates a carte blanche for far-reaching scrutiny of public communication because in practice it is impossible to prove an ad (by definition designed to persuade) is not liable to influence public opinion or voting behaviour. However, failing to “truthfully” declare that an ad falls under the regulation’s vague definition may result in a fine from a national public authority – not an insignificant concern in member states where the freedom of political parties and NGOs is under pressure.
In addition to the scrutiny of public authorities, publishers will also have to check whether advertisers have correctly declared ads as political and provided all the required transparency information. They must do so both on their own initiative (to check whether the declaration and information are manifestly wrong or incomplete) and in response to notifications submitted by users. The regulation’s recitals state publishers should focus on ensuring the correction, completion or removal of the specific information that violates the regulation, and take due regard of freedom of expression. However, the regulation creates few actual binding safeguards that require publishers to ensure political advertisements are not unjustly removed. There is little to stop publishers from taking a cautious approach and interpreting the definition of political advertising broadly, or giving advertisers little time to make corrections. Worse, the regulation incentivises over-removal by potentially incorporating the large fines of the DSA (6% of annual turnover) and GDPR (2% of annual turnover), and giving platforms a 48-hour window to address user notifications in the month before an election. This is particularly problematic for small players that are of no real financial value to platforms.
How should these dangers be addressed in the trilogue?
The vague and broad definition of political advertising risks undercutting the regulation’s overall goal to safeguard democracy. The best way to avoid this risk is to cut issue ads out of the regulation. There is no version of this definition that is not so vague that it inevitably captures political speech in general. This is especially true when platforms and national regulatory authorities have such a large role in applying the definition in practice, as they do under the regulation.
If issue-based ads remain in the regulation’s scope, their definition should be limited and clarified. It must be clear that the entire regulation only applies to paid-for messages. The language in the EP proposal that indicates the definition includes ads that simply influence public opinion on societal issues should be removed. Both the Council and EP have also put forward a list of criteria that are intended to clarify when a message is liable to influence voting behaviour, and when it should be seen as a purely commercial ad. But these norms remain so vague they do little to actually clarify the scope of political ads.
What is needed is language that clarifies the definition of issue-based ads should be applied restrictively, and not capture efforts by civil society or other actors (such as the authors of this blogpost) to contribute to the public debate. One way to limit the regulation’s chilling effect is to clarify that the burden of proof to demonstrate an ad is political falls on publishers and public authorities, and that the requirements to bring such proof are high. Going further, the regulation should include binding language stating that advertisers should be given sufficient time to correct wrong or incomplete information, and that minor transparency issues should not lead to the removal of the ad. This may hamper the regulation’s efforts to limit the impact of disinformation campaigns or lobby groups posing as NGOs. But such safeguards are vital to secure the freedom of expression of civil society and other participants in the public debate.
The broader picture
The process, like other recent ones – such as the Digital Services Act and the proposed EU Media Freedom Act – shows that the EU organisation and policy-making procedures have a hard time when it comes to anticipating the consequences of regulating public communication. This is not surprising. The EU is essentially an economic regulator. That is how it came into being and that is what all the actors are focused on. This is reflected, for example, in the influence of the committees in the European Parliament, where it falls on the Committee on Culture and Education in particular to hold up the flag of freedom of expression. The more the EU enacts rules that affect public communication, the more urgent it becomes to integrate the relevant sets of expertise into EU decision making, strengthen the procedures that anticipate broader impact on the marketplace of ideas, as well as re-think more generally the legitimacy the European Union has for adopting speech-related measures under the legal bases to regulate the internal market and protection of personal data.
Conclusion
The proposed political advertising regulation makes genuinely important contributions. For a long time, the opaque nature of online targeted political advertising has made it impossible for journalists, civil society, or academics to scrutinise who is targeting which voters, with what messages, and how much they spend to do so. Especially the ad library provisions proposed by the EP would bring much-needed transparency. They would provide more granular information on targeting and funding than the DSA requires, force large online platforms to include political ads in real time, and have the Commission run an ad library for political ads that are not published on large online platforms. Similarly, the limits the EP imposes on ad delivery reduce platforms’ power to decide who sees what political ads far beyond what the DSA requires. Nevertheless, it is crucial that the political advertising regulation only addresses political advertising, and leaves the ability of all of us to participate in the public debate unaffected.