Reports that Cambridge Analytica used tens of millions of Facebook profiles to help Donald Trump's 2016 United States presidential election campaign have caused a worldwide furore.
But the ways in which such digital data is put to work – and not just where the information comes from – has been a controversial subject for several years.
The use of artificial intelligence (AI) and big data, in concert with voter information, to create highly individualised advertising messages is now central to political campaigning.
And the online advertising that results from the use of AI and big data (which involves population data, rather than sampling information) is capturing an ever-growing slice of campaign budgets.
The technology’s power lies in its ability to enable campaigns to tailor messages to the particular characteristics of the voter.
Information such as a voter’s age, race, gender, income, interests and much else is factored in by complex algorithms to determine which messages will work best with them. In the US, information from publicly available databases can be particularly revealing, showing a voter’s election participation history and their registered political preference, as well as personal characteristics such as ethnicity. Social media posts can be more useful still.
A person likely to have left-wing or liberal sympathies may receive political advertisements focusing on social welfare or inclusivity. Someone who is more conservative may be the focus of very different messages.
These digital advertisements may link to particular web pages that press home the message, something Mr Trump’s campaign was especially effective at doing, thanks to Cambridge Analytica’s help.
_______________
Read more:
[ Facebook suspends Trump campaign data firm Cambridge Analytica ]
[ UK lawmakers say Trump campaign advisers misled them over use of Facebook data ]
_______________
In particular, some observers have suggested that such AI and big data-driven methods increase the polarisation of the political landscape, with campaigns more narrowly focusing on their likely supporters.
There are also fears that, by producing “micro-targeted” messages, campaigns can say whatever voters want to hear. Different groups may each hear a different “partial manifesto”, undermining informed choice.
Or campaigns may bombard those seen as sympathetic to their opponent with negative messages about that candidate.
Multiple digital platforms are used, ranging from Facebook to Twitter, Instagram, Snapchat and YouTube.
Where once political campaign managers held on every word of pollsters, today they may be just as interested to hear what data experts say – and as a result, consultancies have sprung up offering this expertise. They are especially effective at targeting voters in marginal electoral areas.
Among these specialists is Avantgarde Analytics, which was founded in 2016 by Dr Vyacheslav Polonski to offer "ethical audience targeting". Writing for The Conversation last year, Dr Polonski said there were positive sides to the use of the new technology, saying that it could "be used to run better campaigns in an ethical and legitimate way".
“We can, for example, programme political bots to step in when people share articles that contain known misinformation,” he wrote, adding that these could issue warnings about such suspect information.
Similarly, big-data techniques have been used by activists to track digital advertising, enabling a better understanding of what influence it might be having.
