Act responsibly during election season, online platforms urged

Act responsibly during election season, online platforms urged

USM's M Selvakumar says tech companies must do their part to prevent widespread polarisation online, including through AI-generated and 3R content.

mobile phone socmed
Media platforms are advised to introduce measures to curb the spread of misinformation on social media, messaging platforms and other online channels in the lead up to the Nov 29 polls in Sabah. (Freepik pic)
PETALING JAYA:
With the Sabah state election campaign in full swing, an academic has urged social media and messaging platforms to exercise greater oversight to curb the spread of polarising or sensitive content.

Although the Online Safety Act will only take effect on Jan 1, Universiti Sains Malaysia’s M Selvakumar said online platforms must play their part in curbing widespread polarisation on social media.

He said elections are always marked by disinformation, from misleading statistics and subtle insinuations to coordinated cybertrooper campaigns and inflammatory content leading up to polling day.

Selvakumar, director of USM’s Cybersecurity Research Centre, explained that social media algorithms reward strong emotional reactions, driving polarising and sensational content to the forefront and trapping voters in echo chambers that reinforce existing biases.

He said Malaysia’s digital landscape is further complicated by a “private viral loop”, where racial and religious messages spread rapidly through WhatsApp and Telegram groups, escaping the scrutiny of fact‑checkers and authorities.

The accessibility of generative artificial intelligence adds another layer of risk, he told FMT.

“We may see convincing deep fakes or AI-generated imagery pushed out 24 to 48 hours before polling day,” he said, adding that the aim is to inflict “maximum damage” on the target before such content can be debunked.

In the past three years, the Malaysian Communications and Multimedia Commission has asked social media platforms to remove tens of thousands of problematic posts, including almost 8,000 touching on race, religion and royalty (3Rs).

Selvakumar said that in recent times, the final 72 hours before polling day have been the most “dangerous”, with inflammatory 3R content often pushed to stoke fear among voters.

He said this last‑minute surge is often dominated by short‑form videos, which are harder for automated systems to detect and far easier to make viral.

“It is a well-established pattern seen not only in Malaysia but around the world, and it becomes more sophisticated with every cycle,” he said.

Selvakumar said the upcoming Online Safety Act would mark a major shift in regulating Malaysia’s digital space, replacing laws that were largely reactive in enforcement and focused mainly on individual users.

“Only after a report is made would authorities act. Platforms were essentially treated as neutral intermediaries. The new Act flips that model,” he said.

The Act creates a standing, systems-based framework for safer online spaces. Gazetted on May 22, the law comes into force on Jan 1, with authorities now finalising its detailed rules.

The government also plans to enact ten subsidiary laws and set up an Online Safety Committee to oversee enforcement and industry obligations, as outlined under the Act.

Selvakumar said the law places a statutory duty of care on platforms, requiring them to be locally licensed, equipped with systems to curb harmful content, and obliged to intervene proactively.

“The penalties are impactful, and the message is clear: tech giants can no longer profit from engagement driven by hate speech or disinformation while disclaiming responsibility for the environment they create,” he said.

While most service providers are on board with the new law, he said, some have raised jurisdictional arguments, with US-based firms in particular claiming to operate under their home country’s free-speech traditions and likely to resist local licensing requirements.

“It is strategic. Accepting strict liability in Malaysia could set a precedent that other countries might follow,” he said, adding that the revenues of major tech firms may be affected.

He said that in a heated political climate, a false rumour could trigger a real-life confrontation within an hour. Online platforms must act as circuit breakers that stop online narratives from escalating into offline harm.

Selvakumar said the groups most vulnerable to unchecked algorithms are seniors, who are often targeted with AI‑generated material, and youths, who risk being drawn into harmful content.

“These systems monetise engagement, and anger and fear generate the most engagement,” he said.

A lack of regulation would incentivise amplifying conflict, he said, while the new law forces platforms to align their profit motives with the safety and stability of Malaysian society.

Stay current - Follow FMT on WhatsApp, Google news and Telegram

Subscribe to our newsletter and get news delivered to your mailbox.