Analysts: Hate Speech Normalized During Malaysia Elections
Based in the capital, Kuala Lumpur, the Centre for Independent Journalism (CIJ) set up teams to monitor social media sites over four weeks.
Focusing on key areas of race, religion, royalty, gender, sexual orientation, and migration, the CIJ teams spent hours poring over activity from politicians, political parties, media outlets, government agencies and key opinion leaders.
Their preliminary findings: Hate speech has become "normalized."
One aspect the researchers saw consistently was offensive or discriminatory language being used, said CIJ Executive Director Wathshlah Naidu.
"That is very worrying because we are now normalizing it, and by normalizing it there is a huge potential for it to escalate," Naidu told VOA.
The month-long monitoring project "Say No To Hate Speech" began on July 24 as six Malaysian states geared up for elections.
The vote was seen as a referendum on newly elected Prime Minister Anwar Ibrahim and the strength of Malaysia's Islamic opposition.
Anwar was appointed prime minister in November's general elections and after a hung parliament formed a unity government.
During that election and others, the CIJ had seen a surge in hate speech and disinformation.
"We saw a pattern in the last few elections, where race and religion is often used as a tool," Naidu said. "What we decided was to see how these different issues basically are used to influence voters to see to what extent does the narrative escalate to hate speech."
Monitoring and measuring
To do that, the center collaborated with the University of Nottingham Malaysia, Universiti Malaysia Sabah and Universiti Sains Malaysia, involving dozens of monitors.
"We got people to come into the office, in shifts of two. All day, 12 hours of monitoring. And we had a team of 26 monitors over the weekdays and then another set of monitors over the weekend that was 20-plus monitors," Naidu said.
The teams measured the severity of hate speech on four levels through social media platforms including Facebook, Twitter, TikTok and YouTube.
The levels ranged from speech defined as disagreements and non-offensive language to language or content that could be deemed as incitement.
The project used artificial intelligence developers to track hundreds of keywords and scrape data. But when it came to defining what constituted hate speech, it fell to the monitors.
"The most important thing they do is they are the ones who take the severity level. That requires a lot of human touch," Naidu explained.
And if the team detected anything of concern, their response would always depend on the level of the issue.
"The purpose of the rapid response is actually not really focused on countering anything but making sure that you're able to reduce the impact," said Naidu. "For example, when we see a racist post coming up, and we see people online who are affected, we reach out to a couple of them to see how they're doing, and then talk to them and try to understand their perspective."
The center established the rapid response system with its partners, including Architects of Diversity, Beyond Borders, Justice for Sisters, KRYSS Network, North South Initiative, Sisters in Islam, Pusat Komas and Persatuan Sahabat Wanita.
The monitors also flagged hate speech incidents with the social media platforms.
With a population of more than 33 million, Malaysia is predominantly Muslim.
It also has a history of political divisiveness, says Niki Cheong, a lecturer in digital culture and society at Kings College in London who researches online disinformation.
"Hate speech is fairly prominent in Malaysian politics ... of course, what qualifies as hate speech is often debated. That's why projects like CIJ's monitoring project is useful because they break down the different types and level of 'severity,'" said Cheong, a former journalist.
"Political rhetoric in Malaysia has a history of being divisive, and this tends to be more pronounced in anticipation of elections, even before any dissolution parliament or state legislature," he told VOA via email.
Addressing hate speech
Naidu said that the center's findings point to a "normalization" of attacks and a lack of accountability.
"People online find some level of anonymity and comfort and being able to say these things with impunity, because very often, they are not necessarily censured by the other followers. So that's our biggest concern."
As the team works through the findings, Naidu also said CIJ plans to share the research with Malaysia's government.
"We've submitted position papers to how can they deal with a speech right away," Naidu said. "There has to be more strategic change at a very practical policy level."
But Cheong has called for more public access.
"It is important that there is historical record of hate speech, especially in the context of democratic practice. And the report will be useful for various groups including the media, civil society and hopefully, politicians themselves," he said.
"It is one thing to get the public interested in why this sort of monitoring is important, it's another to work with them on how they themselves can navigate and perhaps call out or take seriously such practices," he said.