by Centre for Independent Journalism in partnership with University of Nottingham Malaysia, Universiti Sains Malaysia and Universiti Malaysia Sabah
After the historic 2018 general election, Malaysia was suddenly thrown into political uncertainty in early 2020 following the defections of several government MPs.
This resulted in three prime ministers leading three governments after the 2018 election. With so much political instability, the stakes for the 2022 general election were high, and social media inevitably became an important space for power brokers and politicians.
Social media is a popular space for Malaysians to interact with each other. Since the 2008 general election, politicians, political parties and their supporters have increasingly turned to social media to engage with netizens. It has since become an important channel to shape political discourse, with politicians and opinion leaders bypassing the mainstream and online media to post comments and statements directly on various platforms.
In the run-up to previous elections, online users employed divisive language and hate-based narratives around race, religion and royalty, popularly known as the ‘three Rs’.
- Sign up for Aliran's free daily email updates or weekly newsletters or both
- Make a one-off donation to Persatuan Aliran Kesedaran Negara, CIMB a/c 8004240948
- Make a pledge or schedule an auto donation to Aliran every month or every quarter
- Become an Aliran member
The 2022 general election was expected to be no different. This was especially because a couple of issues had made social media an important space that shapes political discourse, including by promoting hate and toxic rhetoric.
One was the political uncertainty of the past five years. This meant the stakes were high in this election.
The other was the rollout of automatic registration for voters above 18. This meant there were 1.4 million voters aged between 18 to 20, whose voting patterns were unknown but whose use of social media was well-documented.
Hence, the Centre for Independent Journalism (CIJ), in collaboration with the University of Nottingham Malaysia (UNM), Universiti Sains Malaysia (USM) and Universiti Malaysia Sabah decided to conduct a social media monitoring project before, during and after the 2022 general election.
A key objective of this monitoring was to identify the severity levels of hate speech and expose the sources and patterns of hateful and toxic posts on social media, specifically Facebook, Twitter, YouTube and TikTok.
While the politicians themselves may not be the originator of hateful posts or false information, they did not curb or censure followers who responded with hate speech and disinformation. For some politicians, their repeated use of pejorative terms against their opponents pointed to them being aware of this pattern of behaviour and the kind of response it generated.
The monitoring looked at levels of hate speech that were distributed by actors such as politicians, political parties, government agencies, media organisations and key opinion leaders around several key topics. These were the three Rs, gender and lesbian, gay, bisexual, transgender, intersex and queers (LGB+) persons, and refugees and migrants. The monitoring also identified potential coordinated inauthentic behaviour such as bots and cybertroopers.
CIJ developed a severity scale in which it categorised the levels of hate speech as follows:
- level 1 – disagreement/non-offensive
- level 2 – offensive/discriminatory
- level 3 – dehumanising/hostile and
- level 4 – incitement/call for action (violence)
The team adapted the thresholds in the Rabat Plan of Action on the prohibition of advocacy of national, racial or religious hatred that constitutes incitement to discrimination, hostility or violence and used the “three Ms” framework of messenger, message and messaging to analyse the levels of hate speech.
Social media was mined using a customised automated tool which scraped data on identified accounts, using a combination of 544 keywords and character embeddings. CIJ-trained human monitors then reviewed, categorised and tagged the data.
The CIJ carried out a pilot study from 16 August to 30 September 2022 to test the tool and review the data in preparation for use during the 2022 general election. After Parliament was dissolved on 10 October 2022, the final monitoring was carried out between 20 October to 26 November.
Key findings include:
1. Pas president Hadi Awang and his party were the biggest amplifiers of race. For example, Hadi’s TikTok feed claiming the DAP was merely using Malay candidates to gain voter traction had 2.5 million engagement counts on TikTok, the highest across all social media platforms. Posts on race were also found to perpetuate disinformation. Further, after election day, content created by young TikTok users to manufacture fear went viral and had high cross-platform amplification. Content on race frequently intersected religion
2. The monitoring saw a shift in focus where religious narratives became the main divisive and polarising tool, as it clearly intersected with race and became the basis for the attacks on LGB+ persons. There was high cross-platform amplification, including through posts by key opinion leaders and other influencers. One was by famous singer Jamal Abdillah, who called on voters to reject the DAP which if voted into power, would purportedly restrict the azan (call to prayer) and tahfiz (religious) schools. And the other was by actor Zul Huzaimy Marzuki whose wish to “slaughter infidels” was only prevented because of the law
3. The targeting of refugees and migrants on social media did not seem to be part of any coordinated political campaign. However, it ranked highest for severity at “level 4”. Posts about these two communities contained explicit suggestions calling for physical harm, damage or death. The dehumanisation of these two marginalised groups was made worse when the Immigration Department asked social media users to complain and submit information about people they suspected were without documents. It is however noted that the attacks against refugees and migrants on social media have been consistent and in fact continued post elections
4. Hate speech against the LGBT+ community was, by contrast, clearly a political tool, often coupled with terms such as “liberal” and “anti-Islam”, and used largely to discredit Pakatan Harapan, the DAP and Anwar Ibrahim. At the same time, body shaming and the policing of modesty were aimed at women from across the political spectrum
5. Posts on royalty, which were low during much of the monitoring period, spiked after polling day when it became clear the institution would be a key player in deciding who the next government would be. Most posts under royalty were not attacks against the royal institution but against other actors. One example was the attack on Mahiaddin Yasin for “disloyalty” to the king in rejecting the royal advice to form a unity government
6. The automated system and monitors identified coordinated inauthentic behaviours during the monitoring period. However, these did not have any significant impact on shaping narratives about the issues that were monitored. User-generated comments from regular social media users were the biggest amplifiers.
The monitoring has sparked interest and debate on the key role of the state in combating hate speech on social media.
The virality of simplistic and inaccurate TikTok content, for example the May 13 videos post-election or Hadi’s video accusing the DAP of using Malay candidates has also brought to the forefront the question of social media accountability, especially in relation to the standards and processes they adopt in moderating content.
There is a critical need to interrogate the use of social media as a key tool in shaping the political and electoral discourse and what further standards and measures are needed moving forward within this technology-driven environment. – Executive summary of full report: “Social media monitoring of Malaysia’s 15th general elections“/CIJ and partners