Data in Politics Is an Asset, Not a Threat to Democracy
Tim Berners-Lee—inventor of the World Wide Web—recently marked its 28th birthday with an open letter describing trends online that concern him, including the use of data-driven campaigns to target voters with personalized social media ads that lie to them or intentionally deter them from voting. But data-driven campaigning is not the cause of these problems. Lies and dirty tricks are not new in politics, and their appearance online should surprise no one. Moreover, data-driven campaigning can be used to improve participation and electoral competition. So while it should not be exempt from normal election rules, neither should it be treated as a special case or a problem in itself.
Even in small countries, political activists require tremendous amounts of money and volunteers to mount a nationwide campaign with traditional tools, such as leafleting or direct mailing. Data-driven campaigning can bolster political participation and electoral competition by making it easier for small political groups to organize and grow and by helping campaigners to reach out directly to particular people, as well as disinterested and disillusioned non-voters.
The ability to target personalized ads allows campaigners to reach out to people who do not normally vote. Barack Obama’s 2012 reelection campaign made significant use of data to target nonvoters and encourage them to go to the polls, helping him to win on the support of a diverse and eclectic coalition of voters. Targeting allows rival campaigns to compete for every individual vote, instead of relying on core support bases grounded in class or sectarianism, or competing for focus group archetypes of the center-ground swing voter.
Tim Berners-Lee overlooks these good uses of data-driven campaigning when he highlights their obvious corollary: “unethical” use of targeted ads to discourage particular voters. He also ignores the fact that willful attempts to discourage people from voting are not new and not grounded in new technology. Negative campaigning serves precisely this purpose, and it can reduce overall turnout. Politicians who practice it are well aware of its effects, as evidenced by the tendency of attack ads to focus more on the opponent than the candidate, and by those politicians who go as far as to illegally masquerade as other political groups in an effort to avoid blowback from attacking their opponents.
Mr. Berners-Lee asks, rhetorically: “Targeted advertising allows a campaign to say completely different, possibly conflicting things to different groups. Is that democratic?” The implied argument pulls a thread of causality from new technology, to old campaign tricks, to dynamic voter behavior. One could call it technological determinism. In any case, it does not make sense. Even if dirty tactics were the result of new technologies—and they are not—the Internet still gives citizens unprecedented power to educate themselves and decide how to vote. Not even the most effective targeted ads take this power away from them. Ergo, targeted ads that are discouraging, but which break no laws, are no more unethical than the well-documented dark art of negative or dishonest campaigning (notwithstanding the fact that positivity and honesty are sometimes mutually exclusive).
None of this is to say that data-driven electioneering should not be subject to the same legal standards as all other political campaigning. For example, British law states that all election materials and canvassers must identify the campaign. This allows voters to know who is responsible, and prevents anonymous off-books campaigns that might circumvent spending limits. The origin of political ads on social media should also be identifiable for the same reason, including when they direct readers to third-party content. Similarly, campaigns should not be able to disguise their ads through unofficial collusion with third parties, and regulators should pay close attention when other actors’ digital campaigns might constitute in-kind donations that have been spent, lest they confer an unfair advantage that breaks the spirit of the rules.
But there is a difference between measures such as these, which ensure existing standards remain effective and free from loopholes, and treating political activity on social media as something dangerous in itself that must be tamed by government. The latter represents a dangerous inclination among some policymakers, such as those in Germany who wish to censor so-called “fake news” on social media, operating on the assumption that citizens cannot be trusted to judge what they read for themselves.
In his letter, Mr. Berners-Lee says he opposes such censorious measures. But by making the mistake of viewing degradation in political discourse through a technological lens, he concedes far too much of the argument to the censors. Social media and data analytics are valuable tools for political activity, which means they are inevitably at the disposal of those who say things that are not true or fair, along with everyone else. As Saul Bellow remarks in the opening lines of The Adventures of Augie March, “there is no fineness or accuracy of suppression; if you hold down one, you hold down the adjoining.” Attempts to stop dishonest social media campaigns risk undermining the honest ones too, especially if policymakers ignore the latter while blaming the former on the technology itself.
Image: Jarle Naustvik.