What is Political Microtargeting and what are the risks involved?
Advancements in technology have impacted many sectors including the political world. Unlike the traditional means of conducting campaigns, technology has now made it possible to conduct data driven campaigns in large scale and with high levels of specificity. On account of increased use of digital tools and channels of communication, individuals are now leaving behind digital footprints with vast amounts of data that can be used to make inferences about an individual or group. For political parties, this has translated into greater clarity on the preferences of the electorate through hyper-individualised communication in a process known as micro-targeting.1
The practice of political micro-targeting, although not new, has grown in scale and attracted a great deal of attention for two reasons: the emergence of social media as a communication channel and the existence of big data.2 Micro-targeting is a multi-step process that involves collection of data to analyse it with the aim of understanding people’s behaviour and opinions.3 Collection of data is followed by categorising individuals based on their inclinations such as similar concerns and opinions over matters.4 This often involves analysis of large data sets and use of predictive modelling that matches an individual’s personal characteristics with their political beliefs so as to produce a desired voting decision from that individual.5Messages are later disseminated to the relevant audience.
Micro-targeting has attracted its fair share of criticism due to its recorded harmful effects on privacy and democratic values of a country.6 However, as there is little data on the degree to which the different uses of data characterise different campaigns it is difficult to determine the extent of a practice and whether it is problematic.7 While the true impact of micro-targeting is yet to be seen8 in many countries other than a few that have been documented in the recent past, the effect of micro-targeting in data-driven campaigns cannot be understated.
The most famous example of political micro-targeting is the Cambridge Analytica scandal which was exposed by a whistle-blower. The British data analytics firm allegedly deployed psychological profiling based on social media data to predict and influence voter decisions.9 The exposé, tagged the Cambridge Analytica-Facebook scandal, revealed that the firm had obtained data from an approximated 87 million Facebook users via a third party app and created psychographic profiles on them for political micro-targeting.10 Specifically, in 2013, a ‘Big-Five’ personality test was circulated by Analytica via an app which had participants agreeing to share their Facebook data through the app for academic use.11 These included their identities, addresses, friend networks, and “likes”.12 Though Facebook permitted app developers to collect data from users’ friends, it prohibited sharing this data with third parties.13
Online Political Micro targeting
Political microtargeting has been made possible and easy especially because of the extensive use of social media. In discussing political microtargeting which mostly occurs online, it is important to understand what the phenomenon entails. Online political micro targeting involves, ‘creating finely honed messages targeted at narrow categories of voters based on data analysis ‘garnered from individuals’ demographic characteristics and consumer and lifestyle habits.’14
It can be through political direct marketing whereby ‘political actors target personalized messages to individual voters by applying predictive modelling techniques to massive troves of voter data.’15 It can also be seen as ‘a type of behavioural advertising namely political behavioural advertising.’16 Behavioural advertising entails tracking people’s behaviour online and using the collected information to ‘display individually targeted advertisements.’17
Online political microtargeting can be used in two ways namely:
For identifying voters who are likely to vote for a certain party thus targeting them with mobilizing messages.18
It enables a political party to ‘select policy stances that match the interests of the targeted voter.’19
Online political microtargeting in Kenya
In Kenya, a former executive to a British data analytics firm is on record stating that they rebranded a well known party in the country twice, wrote their manifesto and did research and analysis.20 The data analytics firm is quoted as having said that the surveys conducted covered ‘key national and local political issues, levels of trust in key politicians, voting behaviours/intentions, and preferred information channels’.21 As a result, the company described its operations for the 2013 elections as ‘the largest political research project ever conducted in East Africa’ and further admitted to using tribal divisions in its political messaging.22
The firm is suspected to have used the large-scale data gathered from the aforementioned surveys, Kenya’s publicly available voter registration databases and the data it collected from Facebook to conduct online political micro-targeting on digital platforms to sway voters’ decisions.23 To provide an example, with the requisite data, advertising options on a platform like Facebook can be used to micro-target voters during elections. Some of the ways in which audiences can be segmented and micro-targeted on Facebook are through advertising tools like ‘custom audiences’ and ‘look-alike audiences’.24 Custom audiences allows advertisers to create audience segments that they want to include or exclude in paid political advertisements.25 In doing so, political actors can ‘upload the voter file they have purchased and match other information they have about you to your voting history.’26 Look-alike audiences allows ‘advertisers to upload a list or select a custom audience of people and then, using a complex algorithm, create an audience that is likely to be just as receptive to the messaging as the initial custom audience.’27 Presumably, tools like this built on the psychographic profiles that the firm built from the data it collected.
Risks of online political microtargeting
Invasion of Privacy
Since online political microtargeting involves gathering and combining people’s personal data on a massive scale to identify political preferences, the data gathered threatens the privacy of individuals. For instance if people are suspicious that the websites they visit are being tracked, they may not be comfortable to visit certain websites. Also, by tracking people’s use of the internet, a company can come up with a ‘database of individuals and their interests.’28
Data is prone to cybercrime offences especially if adequate measures are not taken to protect it. Offences where the computer is the target can interfere with data that is already in the computer. The conduct which these offences seek to address include:
Where personal data has been collected for microtargeting purposes and adequate measures have not been put in place to protect the data, hackers or others can access databases containing the personal data and misuse it.
Misuse of personal data
The personal data collected for microtargeting purposes can be used for other purposes which can even be harmful thus threatening the privacy of individuals.
Manipulation of voters
A party could use ‘tailored information that maximises or minimises voter engagement.’32 The targeted information can be false and still have maximum impact. Gorton warns that microtargeting facilitates the spread of misinformation.33
Microtargeting can be used by political parties to exclude certain voter groups. Some groups of voters can be ignored during the campaign season because a political ‘does not expect them to vote’34 or it has high expectations of winning elsewhere. Also certain voters who are deemed not likely to vote can be excluded from receiving political messaging, essentially exempting them from meaningful political discussion.35
As compared to earlier years, technological advancements have led to increased use of social media platforms. This means that political microtargeting online is likely to intensify. As has been illustrated above, the personal data of individuals is likely to be intermeddled with unless appropriate measures are taken to protect the data. As a starting point, since the practice of online political microtargeting is an emerging area, more research needs to be done in order to get more information about the existing benefits and risks so as to educate people.
Also, social media sites can come up with measures to restrict political microtargeting practices. Some social media sites have already taken measures to restrict these practices. For instance in October 2019, the twitter CEO, Jack Dorsey, posted a tweet saying that they had decided to stop political advertising on twitter globally. The purpose of this is so as to protect voters from being influenced. In order to avoid voter manipulation, social media platforms can take similar measures to protect users of their platfroms.
image is from https://www.nextgov.com/analytics-data/2020/05/house-bills-seek-stop-political-ad-microtargeting/165673/
1 IDEA,Digital Microtargeting(19 June 2018)<https://www.idea.int/publications/catalogue/digital-microtargeting> accessed 1 July 2022
2 Orestis Papakyriakopoulos and others, Social media and microtargeting: Political data processing and the consequences for Germany < https://journals.sagepub.com/doi/pdf/10.1177/2053951718811844 > accessed 1 July 2022
3 Frederik J. Zuiderveen Borgesius and others, ‘Online Political Microtargeting: Promises and Threats for Democracy’(2018) 14 (1) Utrecht Law Review 82-96
5 Ira S.Rubinstein, ‘Voter privacy in the age of big data’ < https://deliverypdf.ssrn.com/delivery.php?ID=673064096005020112031018111100031100001024071012061053073118019088121006070114102091011099101107042108110073127022077102088100031034078007004013096005021104122038047037013018117004108106006114099067101000031099075009008103067120119104072097115004&EXT=pdf&INDEX=TRUE > accessed 1 July 2022
6 The Guardian,Leaked: Cambridge Analytica’s blueprint for Trump victory <https://www.theguardian.com/uk-news/2018/mar/23/leaked-cambridge-analyticas-blueprint-for-trump-victory > accessed 1 July 2022
7 Jessica Baldwin-Philippi, The Myths of Data-Driven Campaigning < https://www.tandfonline.com/doi/pdf/10.1080/10584609.2017.1372999?needAccess=true >accessed 1 July 2022
8Balazs Bodo, Natali Helberger and Claes H.de Vreese,’Political micro-targeting: a Manchurian candidate or just a dark horse?’(2017) 6(4)Internet Policy Review 1-13
9 Carole Cadwalladr, ‘I created Steve Bannon’s psychological warfare tool’: Meet the data war whistleblower < http://www.hec.unil.ch/lspinto/Papers%20&%20CV/The%20Cambridge%20Analytica%20Files.pdf > accessed 1 July 2022
10 The New York Times, Facebook Says Cambridge Analytica Harvested Data of Up to 87 Million Users < https://www.nytimes.com/2018/04/04/technology/mark-zuckerberg-testify-congress.html#:~:text=WASHINGTON%20%E2%80%94%20Facebook%20on%20Wednesday%20said,leak%20was%20reported%20last%20month > accessed 1 July 2022
11 Margaret Hu, Cambridge Analytica’s black box < https://journals.sagepub.com/doi/pdf/10.1177/2053951720938091 > accessed 1 July 2022
12 Kevin Granville, Facebook and Cambridge Analytica: What you need to know as fallout widens< https://www.nytimes.com/2018/03/19/technology/facebook-cambridge-analytica-explained.html > accessed 1 July 2022
13 Andy Kroll, “Cloak and Data: The Real Story Behind Cambridge Analytica’s Rise and Fall https://www.motherjones.com/politics/2018/03/cloak-and-data-cambridge-analytica-robert-mercer/ accessed 1 July 2022
14 Borgesius (n 3) 83
22 Justina Crabtree, Here’s how Cambridge Analytica played a dominant role in Kenya’s chaotic 2017 elections<https://www.cnbc.com/2018/03/23/cambridge-analytica-and-its-role-in-kenya-2017-elections.html > accessed 1 July 2022
23 Abdulmalik Sugow A and Isaac Rutenberg, Securing Kenya’s Electoral Integrity: Regulating Personal Data Use (1 October 2021)< https://www.theelephant.info/op-eds/2021/10/01/securing-kenyas-electoral-integrity-regulating-personal-data-use/ > accessed 1 July 2022
24 Meta, How to Use Custom or Lookalike Audiences<https://www.facebook.com/business/help/572787736078838?id=176276233019487 > accessed 1 July 2022
25 Kaili Lambe and Becca Ricks,The basics on microtargeting and political ads on Facebook <https://foundation.mozilla.org/en/blog/basics-microtargeting-and-political-ads-facebook/ > accessed 1 July 2022
28 Borgesius (n 3) 87
29 Jonathan Clough,Principles of Cybercrime (Cambridge University Press 2010)
32 Borgesius (n 3) 87
34 ibid 88
35William Gorton , ‘Manipulating Citizens: How Political Campaigns’ Use of Behavioral Social Science Harms Democracy’(2016) 38(1) New Political Science 61-80