In the era of social media, data mining and artificial intelligence, democracies need to ensure that the core of our political system — elections — is still healthy. In the UK this is now urgent. Whether it is another referendum on Brexit or an early general election before the scheduled vote in 2022, Britain could face significant choices sooner than we think. Later this month we will, unexpectedly, take part in elections to the European Parliament.
Dirty politics, negative campaigning and propaganda have been around throughout history. Politicians and lobbyists do not always play by the rules during elections. At the University of Oxford we have demonstrated several new problems: the speed and scale at which misinformation about politics travels within the UK, the level of foreign direct interference our democracies now face, and the slow pace of change from the technology companies serving up falsehoods at pivotal moments in political life.
We are probably past the point of industry self-regulation. There is a growing consensus about what the platforms need to do to protect our elections. The non-profit Mozilla Foundation has helped set expectations of Facebook, Google and Twitter. They should build tools so the public can track political content — both direct electioneering from candidates and parties, and information from holders of political office.
We should all be able to analyse campaigns launched by known purveyors of political content. Microtargeting tiny groups of voters is a powerful tool in the wrong hands. So the targeting parameters used to determine who sees specific messages should be shared, in order that we can evaluate who is being reached and how much the ads cost.
A government proposal last week, that digital political ads should be traceable to who paid for them, is sensible. But such disclosure rules need to cover all ads.
Right now, not all political actors declare their advertising. Legal guidelines fall short of providing practical definitions of what constitutes a political ad. Social media companies have come under criticism for their own somewhat crude attempts. And there still seem to be secretive groups distributing political ads without disclosing the funding. The most polarising Facebook content comes from political groups that do not even register, research shows.
In part, this is about tracking the flow of money, but it is also about evaluating the ethical behaviour of social media firms. Did the firms allow advertisers to A/B test a message so much that the end result is extremist, polarising content?
The data provided to the public has to be meaningful and useful, and in shape for researchers and regulators. At the moment, Facebook only reveals that an ad may have cost between £0 and £500. The social network offers a political ad archive for the UK, but not all platforms do. Independent researchers face restricted, or negligible, access.
All democracies should demand these changes, but there’s a vital step to take before the next significant UK election. Social media companies should verify and disclose who is buying the ads — all of them — targeted at our citizens. In the US, Facebook and Twitter required political candidates and committees to provide details of their mandatory tax ID, but comparable measures have yet to be introduced across Europe. Several platforms have introduced requirements for political advertisers to provide other forms of ID, but this has been common practice for business ads and can easily be falsified. If the purchaser of political ads is not registered with the Electoral Commission, not on the UK Lobbying Register, or registered as a foreign agent, they should not be allowed to launch a social media campaign.
This need not be wishful thinking: If political actors are behaving badly, we must prevent them from poisoning public life.
The writer is director of the Oxford Internet Institute and chairs the Oxford Commission on Technology and Elections.