How to “Hack” an Election: Red Flags for 2019

6 minute read

Canada has enjoyed relative stability at a time where frantic politics and flawed elections abound among our peer countries. While malicious interference by foreign rivals has garnered the most media attention, the underlying trends fuelling electoral disruptions are technological developments that are pervasive and in use by benign actors as well. The increasing use of technology, big data and artificial intelligence in political campaigning and the public realm more generally, is opening up new challenges for the electoral exercise. This demands reflection on existing institutions and processes in order to keep legitimacy in the electoral process.

To be clear, none of these elections appears to have been truly “hacked” in the sense of a malicious entity manipulating the results of vote-counting machines. It is more about elections having been “gamed”, with micro-targeting and a deepening potential for public opinion management that comes with new technologies having been exploited in such a way as to outsmart the election process. Think “Radio Free Europe”, but with the power and capability that comes with supercomputers, big data and artificial intelligence. And sometimes with less scruple.

The new challenges for elections in the digital age will need to be carefully managed if the Canada is to successfully navigate its upcoming elections and avoid the kinds of electoral disruptions that have become commonplace elsewhere. Yet this depends in part on a realization that Canada’s elections have not evolved in isolation from the technologies that have played a malign role elsewhere. Although Canada has yet to have a scandal of the proportion of Russiagate or Cambridge Analytica, this is not because of any technological limitation or inherent Canadian immunity.

Technology and Voter Targeting

In the lead-up to the 2015 election, the Conservative Party of Canada was criticized for its extensive data collection through its Constituency Information Management System (CIMS) which gave the party a- to that point- unprecedented ability to interact with its supporters. Not to be outdone, the Liberal Party of Canada leapfrogged the Conservatives in the 2015 election by equipping canvassers with iPads so they could mine voter data right from the front door. In both cases, data on voters was aggregated and analysed to provide strategic advice to party HQ, allowing them to adjust electoral strategy to emerging conditions and improve their campaign effectiveness in real time.

In both of these cases, there was nothing untowards about the data being collected, or how it was collected. Yet as most privacy experts will acknowledge, the issue today is less about data collection and more about the data is ultimately used, and to what end. Knowing some basic demographic information is enough to know what kinds of media you consume and therefore the electoral platforms which target you, and in turn the likelihood that you would vote for a certain party. This can lead to all sorts of developments which, while incremental, can come with implications that are troubling.

In the 2011 election, Conservative canvassers were warned not to waste time on “Zoe”, a fictitious 20s or 30s something urban female condo-dweller with graduate-level education. “Zoe” won’t reliably vote Tory, so the best electoral strategy for Conservatives will be to systematically avoid “Zoe” in favour of more receptive electors. Not to pick on the Conservatives, there is little doubt that the Liberals and NDP have similar campaigning practices stemming from big data analytics. The problem is that “Zoe”, and her counterparts in the other parties’ campaigns, will be left increasingly unexposed to any values they do not already espouse leading to greater polarization and partisan information bubbles.

New Elections: Old Ethics

Many of those operating electorally motivated data mining and campaigning strategies are certain to be ill at ease with the relationship between their work, the quality of democracy and the sanctity of the electoral process. The results of segmentation are prevalent daily with increasing polarization and radicalization of the electorate through partisan information bubbles. All parties are hesitant to abstain from these practices given the obvious benefit of increasingly sophisticated micro-targeting and since very little of it is explicitly illegal. Many high profile individuals have no qualms about even the most objectionable uses of big data in elections.

In the era of data science and AI research methods, how else should we expect the rational minds that run political campaigns to adjust their behaviour to a quadrennial, winner-takes-all, multiple-choice, public opinion poll? Potentially millions of data points per individual are being leveraged in order to flip a single other variable; how you feel on election day. It is hard to imagine a more one-sided contest and one which trivializes the democratic process. Certainly, a party may be able to more easily secure a voter’s consent on election day, but what is its value if that consent is not an honest one, or is based on a too fine massaging of public opinion? It is certainly possible for a government to have consent, without legitimacy.

(Re)Constructing Electoral Competition

If one were to select a single driver for this trend, it would be the exponential availability of electorally-relevant data and the ever-decreasing cost and time required to process this data. The new reality of these technological developments is not going away; even if we wanted to undo this part of the digital age, it’s not clear that it would be within the capability of any government. What is needed instead is a fresh approach to elections that is keenly aware of how 21st century technology have changed democracy.

The status quo it is a tragedy of the commons. It is in the interest of none to be the only political actor to shuns technologically-enhanced electoral tactics, especially when the precise nature of the public’s objections are unclear. Yet at this rate, and without intervention by the state to set clear limitations on these practices, will democracy even survive the era of big data? At the very least there must be clear and firm electoral rules that regulate election campaigns in a way that forces them to compete for votes on the grounds of what is constructive for democracy, not simply what will produce a desired electoral result.

For complementary works on the interactions between technology and democracy, see our companion piece on how technology can enhance democratic institutions in Policy Options.