‘Online harms’ to our digital democracy are still being largely ignored
5 ways the digital revolution demands an update to outdated electoral laws.
Towards the end of April, the redoubtable Information Commissioner Elizabeth Denham told the DCMS Select Committee that she was “was surprised and disappointed” that there was not more attention given – in the new Online Harms White Paper – to “what I think is a huge societal harm, which is around electoral interference and the need for more transparency in political advertising”. Her stark criticism reflects growing astonishment that, despite all the scandals that have come to light since 2016, the outdated laws surrounding elections and referendums remain virtually unchanged. As a consequence, we may well go into another election or referendum later this year in which parties and campaigners can continue to behave in ways that jeopardise the openness, fairness, and legitimacy of the vote.
The reason electoral law is so outdated can be explained in two words: digital revolution. As our lives become more digital, so do our politics. Nowhere is this clearer than during election and referendum campaigns. For the 2016 presidential election in the US, candidates spent $1.4 billion on online advertising. In the 2017 UK general election, all parties – and associated campaigners – dedicated much of their time and money to digital campaigning.
Are these digital developments just a natural evolution of political campaign communication – or a revolution? Following Brexit and Trump, are internet and social media are destroying liberal democracy?
Every major change in communication has provoked exaggerated hopes and fears about its political effects. Some of these fears will be proved groundless. For example, there is already evidence that active social media users may be exposed to more sources of news rather than less.
Get one whole story, direct to your inbox every weekday.
But there are five specific areas where democratic protections, built up in Britain over the last 150 years, appear threatened by digital communication and data‐driven campaigning. If left unaddressed, these will necessarily undermine the legitimacy of elections and referendums.
1. Protecting the secrecy of your vote
A secret ballot was introduced in Britain after 1872, in order to prevent intimidation, bribery or treating of voters. But personal data collection and analysis, and the ability to predict whether and how someone will vote, jeopardises the principles of voter confidentiality.
By pulling together the digital traces voters leave online and combining this with surveys and with other sources of offline data (such as consumer habits), campaigns can quickly accumulate detailed profiles of individual voters. Cambridge Analytica famously boasted it had up to 5,000 data points on every American voter. Even with many fewer data points than this, a candidate can know more about each voter than the voter can know about the candidate.
The consequences are manifold, but are seen most visibly in the ways in which voters are mobilized – or deliberately not mobilized. The Trump campaign admitted that in 2016, for example, that it ran three voter‐suppression campaigns in order to depress the turnout of idealistic white liberals (Bernie Sanders supporters), young women (who were angered by the behaviour of Bill Clinton), and the black vote (on the basis of comments Hillary Clinton made about black males in the 1990s).
2. Protecting voters from undue influence
Facebook tools enable undue influence to be exerted on voters. ‘Dark posts’, for example, allowed A/B or multivariate testing of campaign messages that were only visible to the intended recipient. In the ten weeks prior to the 2016 EU referendum, the official leave campaign, Vote Leave, sent over one billion targeted digital ads, the majority via Facebook. During the 2017 UK general election campaign, eight UK parties spent around £3 million directly on Facebook – with the Conservatives alone spending £2.1 million.
Since then the technology platforms – Facebook included – have introduced political advertising archives, but these only address part of the problem. Many ads are not captured in the archive and much of the crucial associated information – such as who an ad was targeted at – is missing. On top of which, ads represent only one aspect of undue influence online.
3. Maintaining a level playing field between different candidates
In the UK, democratic protection to ensure fairness between candidates developed during the nineteenth century, culminating in the landmark 1883 Corrupt and Illegal Practices Act. This Act, on which successive subsequent UK election legislation has been based, sought to ensure that “the length of a man's purse will not, as now, be such an important factor [in electoral contests]”. In contrast to other democracies, such as the US, the UK has managed to maintain this commitment to fair and equal access by constraining campaign spending, particularly at a local level.
The distinction between the local and the national campaign has never been absolute, but used to be relatively distinct. But since central campaigns have shifted online, parties can use national budgets to wage local campaigns, making the constituency spending limit moot, and privileging the bigger and more well‐resourced parties.
4. Preventing elections from being bought
According to The Economist, “The world’s most valuable resource is no longer oil, but data”. This means that without being able to put a value on the collection, analysis and use of personal data, many of the spending caps on elections and referendums will look increasingly anachronistic.
Some of the value in data is straightforward to calculate in monetary terms. Buying data from a third party, for example, has a price tag. However, a lot of the value of data is much harder to assess. Data can be stored and analysed outside the country in which it is being used, similar to the way tax havens like the Cayman Islands work. High-value algorithms and election software can be gifted to a campaign, or licensed at preferential rates. The Cambridge Analytica files are a vivid example of this.
5. Assessing the credibility of political claims
The same 1883 Act which limited election spending also stipulated that every piece of campaign literature have a clear indication of who it was sent on behalf of, and who published it. This was so that members of the public would know when they were receiving propaganda from a particular candidate, to provide some accountability for election literature, and to discourage campaigns from making misleading statements. The same stipulation does not apply to digital material. A voter can therefore receive a political advertisement in their Facebook or Twitter feed and have no way of telling who it has come from or who has paid for it.
What can be done?
Some of these issues can be dealt with by changes to electoral legislation. Others could be partially resolved by new legislation requiring greater transparency. Yet preventing campaigns from targeting susceptible voters with tailored messages would be much harder.
These issues all raise questions not just about how campaigns are run, but about how we participate in politics, how we are represented, how we discover and verify our political news and information, and whether the digital environment supports or obstructs democratic deliberation. The revolution in communication forces us to rethink the way existing democracy works. Either that, or we watch its legitimacy progressively eroded.
Get our weekly email