Political elections are fought all over the world when candidates knock on doors to canvass, make speeches and debate issues with each other.
Elections are also fought when candidates spread messages about their manifestos on posters – for example, on bus shelters or billboards; and on radio or television.
In each of these the candidates can be challenged on what they say and voters get to form opinions based on what they see and hear.
Since the re-election of Barack Obama in 2012, elections are also now fought on social media, but this has been a much darker business. Cambridge Analytica helped target people on Facebook with personalised advertisements based on predicting personalities from online behaviour in both the US presidential election and the UK referendum on European Union membership.
Part of the reason for this new-found openness is to redress what we now know about what happened in the past
There is nothing actually wrong with what they did, except that the model used to predict personalities was based on data illegally gathered from user profiles of millions of users. That particular loophole has been addressed and in theory it should not happen again.
The upshot of this is that all kinds of elections and referendums now reflect the effectiveness of advertising on social media, because it can be highly personalised and targeted.
Advertising spend
Many have called for election candidates and parties to publicly declare their advertising spend on social media, just as they declare their spending on other outlets. In fact the social-media companies are already starting to do this for them. Since March, Facebook has had a publicly accessible and searchable report on all active advertisements, who is placing them and how much they are spending.
In Ireland, the biggest political spender on Facebook recently has been the Referendum Commission, spending €91,507 between March and June this year on 29 different targeted advertisements, most of them video adverts. For each of the adverts anybody can go to the Facebook report, play the advert, see approximately how many times it was viewed and how much money was spent on placing it online.
As a result of the Facebook report we now know that the Trump Make America Great Again committee, one of his re-election agencies, has been spending more than $1 million per week on Facebook alone with 129,740 different adverts, far more than any other candidate, and that was before his re-election campaign was officially launched last week.
One of these adverts was a 30-second video sent to Facebook users who had not up to that point sent the president an online happy birthday message on June 13th – offering them the chance to do so.
The way the BAI operates is to wait for complaints from citizens and then to investigate them
The active advertisements report is part of Facebook’s efforts to increase transparency in advertising and, just like candidate debates or speeches or public advertisements, we now know what candidates are saying in those personalised, targeted Facebook advertisements.
Part of the reason for this new-found openness is to redress what we now know about what happened in the past. In November 2018, Facebook was forced to reveal a cache of adverts run on behalf of the Vote Leave, BrexitCentral/BeLeave and DUP Vote to Leave campaigns ahead of the UK’s referendum on leaving the EU in 2016. These were submitted to a committee of the UK’s digital media, culture and sports department, which was conducting an inquiry into fake news.
So when we hear calls for regulation to be brought into how social media is used in elections and referendums, it is not just about the amount of money they spend, or the mechanisms used to target people: it is about the content itself, the messages sent to individuals that nobody except the recipient would normally see.
So has the Facebook’s active adverts report shone a light on to the darkness introduced by the use of social media during election and referendum campaigns, and will we ever have another scandal like Cambridge Analytica? Things are better now than they used to be in that we can see the adverts, but it is left to investigative journalists or concerned citizens to monitor them.
Is it possible to have regulation and enforcement in an area as quickly moving as online advertising? Should we even bother trying?
There is a proposal to the Government that the Broadcasting Authority of Ireland (BAI) will police video content on Facebook, YouTube and Twitter from September 2020. This would include enforcing EU rules similar to the way the Irish Data Protection Commissioner enforces EU rules on data privacy.
Complaints process
The way the BAI operates is to wait for complaints from citizens and then to investigate them. The problem here is that when people are shown toxic material on social media, personalised and targeted videos based on their personalities, demographics and what they want to hear, they are not inclined to lodge a complaint to the BAI or anyone else.
Having a robust complaints-resolution system just won’t be good enough to catch the mild form of radicalisation that went on in recent elections and referendums, which will surely continue and increase in future ones.
Instead, the BAI needs to be proactive, to work with and to develop its own artificial intelligence in order to monitor the AI used in advertising on social media. Putting the onus on social-media companies to hire independent agents to investigate complaints from citizens, which is just a form of self-regulation, will not address how social media has been abused in elections and referendums in the past and will keep regulation at least one step behind, instead of ahead of, social-media companies.
Alan Smeaton is professor of computing at Dublin City University