This essay is concerned with the regulation of political advertising online in order to ensure the integrity of elections. It asks what political and digital media entities will need to be considered and the questions that regulation will need to address. This report highlights the continued lack of clarity regarding the scale and nature of online political advertising. Recent elections and referenda have been undermined by the large-scale manipulation of the digital environment, where most voters now get their news and political information.
Given the examples of elections and referendums marred by questions of disinformation and manipulation, a range of digital media and advertising companies have committed to the Code of Practice on Disinformation. Each have submitted regular reports on efforts to address issues such as bots, transparency in funding digital campaigns, and the promotion of problematic content. This report and the wider work of ERGA as well as reports by Mozilla and the Office of the French Ambassador for Digital Affairs offer more clarity as to what specific information is required by election monitors to ensure they can appropriately investigate to uphold the public interest.
This report clearly shows the need for enhanced transparency from digital media companies. One of the core challenges of this, and related research, was that it was limited to analysing only the adverts the companies defined as political. It was not possible to systematically analyse a complete database of adverts to establish what may have been omitted and to understand why. This is a critical blind spot in the transparency efforts by digital media companies. While the political advertising archives help address some issues, they leave others unaddressed and point to new areas in need of further investigation.
The time has come to initiate more focused discussions about what this regulation should look like and to establish what information we should compel political organisations and social media companies to disclose. While some of the social media companies offer some information on political advertising, Irish national regulatory structures could address a wide range of issues. This and be achieved by requesting political parties, campaigners and donors to disclose detailed information on how they advertise online. Moreover, the evident challenges posed by the varying approaches to classifying political and issue-based advertising by just three social media companies underlines the need for a standardised approached across social media companies if election media monitoring is to be achievable.
Given the ease of creativity afforded by digital software, one of the first challenges that any national regulators will have to consider is what exactly constitutes political advertising online? In the analogue era, it was easier to define. For example, the Broadcasting Authority of Ireland (BAI) have clear guidelines26 on what constitutes a political broadcast. But in the digital era paying for political advertising can take on many forms, such as paying companies or employing ad content specialists, to paying for editing an image or bots to liking a post on website like ‘Fivrr’.
Facebook and Google between them own a substantial number of the top ten social networks in Ireland, but ads were placed on other platforms such as Spotify. And there are a range of other platforms where ads can be placed and more will no doubt emerge and many social networks have not agreed to the Code of Practice on Disinformation. This indicates another structural blindspot. European-wide voluntary regulation would need to be supported by national measures compelling political organisations to be appropriately transparent.
The responsibility to uphold the integrity of elections and choose who gets to participate in political campaigns via these platforms lies with governments. But states arbitrating on what should be available in the digital public sphere is highly problematic. On one hand, such measures leave too much authority in the hands of the state with implications for democracy. And on the other, the slow instructional process of legislation and policy development makes is undesirable. However, a powerful arbiter is one that is independent of both state and private entity, that can mediate between the two while holding both to account and can act in the public interest.
Addressing social media and providing more information to ERGA
Neither Google nor Twitter labelled issue-based advertising at all over the election period, and their political advertising archives did not offer enough insight to ensure electoral integrity in this regard. Given that voting patterns can indicate the centralisation of certain issues, such as environmental issues or immigration, it is necessary for countries to establish who is buying advantage for their political messages. Issue-based adverts were most likely to run without a disclaimer and without the payer information on Facebook. This indicates that these types of adverts are the most vulnerable to coordinated manipulation. In the current format, it is not possible to establish if online advertising is being used to manipulate public perceptions of issues, or political parties and candidates that do or do not support such issues. This is precisely what we need to know.
We also found that Facebook, Google and Twitter tend to present some information in aggregated form or as a summary when what is required is more detail. For example, spending information tends to be reported as a total over a period, rather than spend per advert, which would facilitate a better understanding of increases and decreases in spends before and during elections. To meet the commitment to ensure electoral integrity requires improvement on the degree of transparency on a number of areas. All three platforms should provide information in such a way that allows election monitors to easily establish the volume of advertising over a defined period. This includes detailed information on the pages that adverts appear on, who funds them, and the issue-based adverts labelled with the specific issue of importance. Additionally, this information should be provided in a format that is easily analysed from a number of perspectives.
What we can do nationally to supplement EU codes?
While Facebook, Google and Twitter have made some efforts to enhance transparency regarding political advertising, Irish national regulators and political parties could do substantially more to codify and produce suitable information regarding their production, purchase and distribution of advertising
online. In the case of the European elections in Ireland, this report found that often the sponsors of adverts were primarily political parties or candidates. However, there were many examples of names and entities paying for political adverts that had no apparent connection to the beneficiary. This poses a question for regulation to consider: who should be allowed to pay for advertising?
The production and services required to effectively produce high quality content for social media and the investment in services to enhance the reach of adverts across platforms can be substantive. Regulation needs to consider whether there should be a requirement for political organisations to disclose the spending on content production and distribution. There is an urgent need for the Standards in Public Office Commission to set clear guidelines regarding who may pay for advertising (party candidates, official accountants, communications managers or any party member?) and the disclosure of names and amounts spent on content creation and distribution.
The Irish national regulator, where possible, should require more detailed disclosure of a political party’s or candidate’s campaign spending including a list of all digital and analogue platforms: how much has been spent per advert; for how long was the advertising scheduled to run, and how much was spent on content creation; were micro-targeting options used to either target or exclude demographics; and how many official political party, local branch and candidate pages are on different social media platforms? By requiring more detailed information from political parties, regulation will also facilitate better identification and the ability to address national and international malicious actors.
Ireland is a small media market where it is challenging to sustain individual and independent monitoring initiatives. Considering this, the best approach may be through collaborative networks of interdisciplinary monitors, researchers, fact-checkers, investigators, and open-source intelligence (OSINT) groups with a public face that can engage in best practice for public interest communication. Other markets – for example, in the UK, USA and across Europe – have a number of public interest institutions (such as NGO initiatives, independent fact-checking and verification organisations, media monitoring and research groups) already in operation. In the UK the Digital, Culture Media and Sports Committee’s report into Disinformation and Fake News27 was pivotal in securing further transparency from Facebook regarding political advertising during the Brexit Referendum in 2016. Such initiatives are regarded as vital public resources and essential part of the digital infrastructure required for democratic societies. However, recognising that this is a smaller media market, to tackle problems that arise in the information ecology, pooling expertise and skills in an independent project can be of benefit.
The desire and ability to evade the regulation or measures introduced by social media companies limits the effectiveness of such efforts. Where regulation presents a barrier, malicious actors develop ways to overcome them. Something that can respond to new problems in a meaningful way must be implemented. To address the fast evolution of evasion techniques, and to monitor the practices of malicious actors, a range of safety nets that uphold the quality of the information environment should be developed. Ad hoc volunteer efforts cannot achieve what is required to comprehensively address the range of problems that are developing in the digital environment. In Ireland, there is a pressing need to develop a media monitor that can combine the skills and expertise needed to address these problems such as OSINT, fact-checking and verification source tracking, image and video manipulation expertise, emerging technologies such as blockchain. For too long now research and industry has recognised that regulation is needed, it is high time to start discussing what specifically this might look like and ensure that it, above all, works best for the public.