If Prime Minister Datuk Seri Anwar Ibrahim has his way, Malaysians won't be using Waze any more. The government has directed civil servants to stop using the navigation app, citing concerns about data security and the company's country of origin.
"This is not merely about the Waze application but its connection with a company headquartered in Israel," he said.
Personally, I thought this was a strange issue to raise out of the blue. On the other hand, a short chat with a journalist friend revealed that he wasn't particularly surprised by the timing, given that an election is likely to be due within the next year.
Indeed, for all the criticisms the government has faced, one thing the Prime Minister has been praised for is his steadfast opposition to Israel's actions in Gaza, a position that resonates with many Malaysians across the political spectrum.
There were jokes about "ban haze, not Waze", questions raised in Parliament, and all sorts of commentary on social media. Like fireworks in a soundproof room, plenty of noise but little illumination.
The uncomfortable truth is that organised political influence has been masquerading as spontaneous public opinion for years.
In 2016, the Philippines became what researchers famously described as "patient zero" for global disinformation. With relatively little money to spend, Rodrigo Duterte's presidential campaign turned to social media to reach voters.
His campaign created multiple "click armies" to translate and spread his messages. These groups were then encouraged to compete with one another, pushing the boundaries of their tactics and becoming increasingly effective at tapping into the anger and resentment of their target audiences.
His success encouraged others to follow later that year, including the Brexit campaign in Britain that saw the country exit the European Union and Donald Trump's first presidential election victory in the United States.
Indonesia provides another example. Political "buzzers", as these online campaigners are known, illustrate how commercialised such operations have become, with politicians turning to marketers and social media agencies.
"Politics enters that field because a number of studies say that social media conversation and positive sentiment determine someone's electability," said one marketer in a 2017 case study of the Jakarta gubernatorial election.
Malaysia has not been immune. During the 2022 General Election, TikTok became an important campaigning tool, particularly for reaching younger voters following the Undi18 constitutional amendment, which lowered the voting age to 18 and introduced automatic voter registration for millions of young Malaysians.
A study by researchers at Singapore's Nanyang Technological University found that parties deployed content creators, videographers, and influencers to produce a steady stream of political content. Successful creators could reach hundreds of thousands of followers, with videos that attracted millions of views.
And they can be paid handsomely for it. An undercover investigation by the media organisation The Fourth found that a Malaysian cybertrooper operation working on a political campaign could be allocated a budget of up to RM3mil for the six to 12 months leading up to a general election.
But does all this effort actually work? The short answer is "not really".
Research involving Facebook and Instagram users during the 2020 US presidential election found that changing people's feeds, including reducing their exposure to like-minded political content, did not produce measurable changes in their political attitudes or evaluations of candidates. Perhaps because humans are surprisingly stubborn creatures. It takes a lot to change our minds.
Yet that doesn't mean spending time making a seven-second video that you scroll past in five is completely pointless. Repeating the same message over and over keeps candidates, their messages, and their faces in our minds. And when those messages come from close friends, we're more likely to pay attention. The upshot is that this continuous stream can make some difference come election day.
The authors of a 2012 study in the United States, again involving Facebook, found that social messaging on the platform increased voter turnout directly by about 60,000 people and indirectly by another 280,000 through social connections. Together, that amounted to roughly 340,000 additional votes, or about 0.14% of the voting-age population. So this stuff does work in part.
Which brings us back to Waze. I’m not convinced the announcement itself is about attention-getting. Perhaps once the government clearly explains its reasons, the hubbub and murmurs will die down.
But as elections approach, we can expect social media to become noisier. Some of what we encounter will be honest political argument. Some will be organised campaigning. Some may be paid influencers presented as spontaneous public sentiment.
More worryingly, some of it may be generated or amplified by artificial intelligence (AI). Recent experiments published in the journal Science found that conversations with AI models could substantially shift participants' preferences for political candidates. The catch? Efforts to make the models more persuasive also made them more likely to present false information.
Nothing is perfect, be it politicians or the pablum they offer. But we should at least be wary of confusing popularity with consensus, repetition with truth, and a thousand accounts saying the same thing with a thousand independent people reaching the same conclusion.
Perhaps the noise and chaos are simply the price of democracy. The danger begins when we can no longer tell who's shouting, why they're shouting, or whether they even believe in their own lamentations.
