Cambridge Conference on Electoral Democracy

Online Seminar 25 October 2023 - Framing the challenge of social media in elections

​

Summary of Discussions 

 

The online discussion was assisted by presentations from Mauritius, South Africa, Malaysia and IFES. 

 

There is little doubt in the minds of electoral commissioners participating in the seminar that new technology, including artificial intelligence (AI) and the use of social media to spread misinformation and disinformation about electoral processes has significant potential to damage trust in electoral processes and consequently cannot be ignored in the strategic operations of electoral commissions. Electoral commissions, with a mandate to protect electoral integrity, are of the view that specific strategies to counter this potential need to be developed. Identifying just what those strategies are, which institutions and organisations, both governmental and civic, are best placed to implement them and the capacity, skills and resources needed to do so effectively is the purpose of this seminar and the 2024 conferences in Mauritius and Cambridge.

 

Misinformation and disinformation are not new phenomena—political campaigners commonly make claims and counter claims about the motives of their opponents and the risks of choosing “the other side”, and not all of these claims are true. Traditional forms of information dissemination such as print media, TV and “stump speeches” have in many elections been replaced by social media turbo-charged by new mobile technologies. Social media in itself is relatively neutral. However, the characteristics of social media—the manner in which it is being used by political actors to deploy information which is misleading or wrong, it’s ubiquity, immediacy and low cost, the apparent unwillingness of social media platforms to invest in contestable debate about “posted” information and the demographics of the electorate to whom it is directed, particularly youth—make it potentially more damaging to electoral democracy when it is abused.

 

In Mauritius, 72 percent of the 1.2 million population are using social media, with FaceBook (Meta) the most popular. Seventy to eighty percent of web traffic is shared information across platforms. The 2019 election saw both formal and informal actors using social media to promote political claims. Moreover, there appeared little action by social media companies to moderate misleading or incorrect content, seemingly driven by perspectives on the US constitutionally guaranteed First Amendment allowing unfettered “free speech”.

 

The demographics of social media users in South Africa are consistent with other countries. The majority of users are female (51%) and skewed towards younger people 18 to 29 years old. This group also has low rates of electoral enrolment. Twenty five percent of the 43 million internet users are using social media, the majority (56%) in the 18-29 year cohort. WhatsApp is the largest social media platform with a 93% penetration, followed by FaceBook, Instagram, TikTok, Twitter(X) and Messenger. The seminar also heard that globally there are some 4.8 billion social media users, and that between April 2022 and April 2023 150 million new users came on line. A significant concern is the growth of hate speech leading, amongst other things, to greater polarisation of the community.

 

The use of social media has shown extraordinary and exponential growth. The internet first came to Malaysia for example in the late 1990s, with only 2.6% of the population having access. Now, about 97% of the 34 million Malaysians are connected online, 26.8 million being active social media users. As in South Africa, WhatsApp (89.3%) is the most popular social media platform. Malaysia had its first “internet election” in 2008—at the time online news portals and blogs were popular and began to influence the manner in which campaigns were conducted. Now, social media platforms have become the new battlegrounds for political contestants to influence voter behaviour.

 

With most if not all electoral jurisdictions facing similar patterns of growth and penetration, the potential for social media to play a significant role in future elections is clearly evident, and worrying—according to IFES, some four billion people will go to elections in 2024 across multiple jurisdictions including the US, India, UK, South Africa and Indonesia.

 

While participants noted that social media can and should be harnessed by electoral commissions to educate voters about electoral processes, and that, in the words of one of the participants, “this lady is not for turning” (signifying that social media is now an integral part of life and the conduct of all manner of business and government services, including elections), discussion focused on the challenges that social media has brought to the conduct of elections and ensuring levels of trust in electoral outcomes and electoral commissions can be maintained and enhanced. These challenges include (in no particular order):

• incorrect and misleading information, once injected into the public domain, has a life of its own—the speed and breadth of information proliferation is difficult to contain and rebut once it has taken hold in the minds of electors;

• social media platforms, initially assisting to prevent the spread of misleading information, have more recently adopted a hands-off approach. Social media platforms concede that resources once invested in content moderation have been steadily withdrawn—speculation on the motivation ranges from a lack of journalistic standards, a lack of understanding about electoral processes or generating clicks is more profitable than engaging in traditional contestable debate and fact checking. This despite evidence that the platforms have automated capacity to effectively remove large volumes of misleading or inflammatory content;

• even where voluntary codes of good practice by social media companies are in place, the evidence suggests that there is little, if any, genuine attempt to abide by those codes. Earlier positive arrangements between electoral commissions and social media platforms seem, now, difficult to sustain for future elections;

• laws relating to electoral misconduct and breaches have not been modernised to deal with misleading and incorrect social media content; and

• the emergence of AI as a tool for those bent on spreading misleading and incorrect information threatens to create sophisticated content that is incredibly difficult for even the most educated consumers of social media content to distinguish between fact and fiction. An emerging view amongst researchers is that AI’s use by autocratic and authoritarian regimes is a real threat to electoral democracy (https://acrobat.adobe.com/link/review?uri=urn:aaid:scds:US:6c85993c-a215-3ac2-ac68-3a2274ca4278). Along with AI, tactics for spreading misinformation include using “influencers”, seeding seemingly non-political content with propaganda, micro-targeting and hybrid campaigns. Identifying and combatting the entire range of communication avenues being used is likely to be impossible even for the best resourced electoral commission..

 

The seminar provided an opportunity for preliminary exploration of “solutions” to the problem of misleading and incorrect information. Three broad categories of response were identified:

• educate the electorate to “check the source”;

• regulate through reform and modernisation of “truth in political advertising” laws; and

• partner with social media platforms, media monitors, civil society organisations and others.

 

These responses, their relative importance, prospects for short and medium term success and interaction with each other will be explored in more depth during the Mauritius conference [details of the Mauritius Conference will be posted here shortly]. The role of electoral commissions in pursuing these strategies will also be explored during these discussions.