Bethany Shiner, Lecture in Law, Middlesex University, London contributed this piece to our report Reining in the Political ‘Wild West’: Campaign Rules for the 21st Century.
The evolution of data-driven political campaigning has spawned an entire industry that has capitalised personal data for political ends.
The scandal around the deceptive and opaque use of personal data and the global web of connections between political campaigns and corporate interests has exposed an approach towards the electorate that seeks to cajole and steer it, not through open and robust debate but through personalised, localised and private digital advertisements.
Such forms of political communication can be positive and empowering, but can also contain misleading, inaccurate or false information that cannot be easily scrutinised. The ICO’s proposal for a statutory code of practice seeks to promote dialogue between the regulators and the government, and encourages a comprehensive reflection on corporate and political practices.
However, due to the complexity of the issue, any attempts to further regulate political campaigning need to be carefully thought through to avoid being ineffective and having unintended consequences.
The ICO’s proposal for a statutory code of practice is an attempt to change the practices of ‘datafied’ political campaigning through clear, enforceable rules. The statutory code should establish standards in political campaigning and limits on the use of data in politics. What is unclear is how a statutory code of practice will sit alongside section 8(e) of the Data Protection Act 2018 which enables the processing of personal data for activities that “support or promote democratic engagement” such as communicating with electors, campaigning activities, and opinion gathering inside and outside election periods.
Of course, there are numerous actors involved in political campaigning, not just registered political parties and campaign groups but also lobby groups, interest groups, online platforms, individuals, foreign countries and private interests. The statutory code must provide further guidance on how section 8(e) may apply to private organisations paid to process data.
It would also be helpful to clarify how section 8(e) sits alongside the additional provisions applicable to sensitive personal data, which includes political opinions. Specifically, how are the methods of using personal data to reveal or infer sensitive information, such as political views, consistent with the Data Protection Act 2018?
For example, when using data obtained from multiple sources and analysed, political parties did not regard any information inferred from this process as ‘personal data’, a conclusion with which the ICO disagrees.
In short, there is a distinction that needs to be clarified between using personal data for political purposes that is surface-level data processing (such as using the electoral register alongside information submitted into a mailing list to send out political messages), and using personal data that is processed to infer sensitive personal data such as political opinions.
Political parties and campaign groups invest heavily in data, financially and strategically, and there is nothing to stop them from doing so.
There is nothing to prevent the marketisation of data for political purposes. There is also nothing that regulates political communication outside of TV and radio political party broadcasts. Therefore, guidance on how micro-targeting can be consistent with the Data Protection Act 2018 would be welcome, bearing in mind the need for a distinction between when an individual is targeted based on data given freely with explicit consent and when an individual is targeted after the processing of other data sets to infer their political views.
One particular aspect of micro-targeted politics is disinformation, i.e. the content not the method of communication. The government has said it is already tackling disinformation through legislative and non-legislative initiatives. The democratic necessity in protecting freedom of expression means any initiative must not curb free speech.
There is a nuance in that some manipulative tools that seek to shape and engineer political discourse, including amplification, bots, troll farms plus micro-targeting contain misleading and manipulative content – but not false or illegal content. In the battle to fight disinformation, we should not regulate political communication in an expansive or suppressive way. Instead, we should focus on the mechanisms of manipulation, not the content.
There have been suggestions that social media companies and intermediaries work closely with regulators and advise political parties on transparency and accountability when using data to target voters on those platforms. The Digital, Culture, Media and Sport committee has recommended that the Electoral Commission establishes a code for advertising through social media during election periods and considers whether social media campaigning should be restricted during the regulated period to registered political organisations or campaigns.
The Committee has proposed a new category for technology companies which is neither platform or publisher, but something in between that establishes some liability to act against “harmful and illegal content”. Social media platforms are being urged to introduce transparency features – with the ICO and the Electoral Commission being consulted on those features and completing evaluations.
If intermediaries will be expected to monitor political content online (to determine what is harmful or illegal, and whether any restrictions are complied with), very careful thought must be given to how this can be done while preserving freedom of speech and not enforcing rules unfairly or in a discriminatory way.
Further, serious thought must be given to whether such power should be delegated to technology companies. The same digital interventions that can be heralded as promoting democratic engagement, such as Facebook’s ‘Get out the vote’ campaigns, can also be used to suppress democratic engagement or shape democratic discourse opaquely.
Such a role could make it more likely that intermediaries, such as Facebook, will disrupt political campaigns – as it did during the Irish referendum on the Eighth Amendment when, after public pressure, it blocked advertisements that originated from outside of Ireland.
Although this was a legitimate concern, such an intervention is in the gift of intermediaries that make judgement calls as moderators and are not subject to review in the way an administrative decision would be. Facebook’s action in the Irish referendum came late in the campaign cycle and was an unforeseen intervention disadvantaging some campaign groups because it disrupted campaign strategies. Interventions such as this should be predictable, consistent and transparent.
Although technology giants have vowed to self-regulate by taking steps towards greater transparency and better monitoring of electoral interference, it has been repeatedly shown that they are irresponsible and have no more regard for the democratic process than their fluctuating stock market value dictates they should. Facebook’s CEO has repeatedly refused to appear before parliamentary committees and reports have illustrated Facebook’s decision to not act on certain forms of electoral interference.
Regulators have so far encouraged dialogue with intermediaries but this should only be done to the extent of consultation. We cannot afford to dilute or compromise electoral integrity according to the desires of private interests and corporate profit.
Also see Bethany Shiner’s forthcoming article in Public Law journal entitled Big data, small law: how gaps in regulation are affecting political campaigning methods and the need for fundamental reform.
Read Reining in the Political ‘Wild West’: Campaign Rules for the 21st Century