KUALA LUMPUR — Malaysia’s Online Safety Act (ONSA) should expand its focus beyond harmful online content and address the growing influence of artificial intelligence (AI) on users’ thoughts, decisions and behaviour, an AI governance expert has said.
Kamales Lardi, CEO of Lardi & Partner Consulting, said ONSA provided an important foundation in tackling recognised online harms such as harassment, emotional distress, self-harm and violence against children.
However, she said the rapid development of AI had introduced new risks that may not be fully covered by existing regulatory frameworks.
“I think ONSA is a great starting point it’s addressing these harms that we already know and recognise that have existed for a number of years but it is primarily content centric.”
“I think AI it introduces a new layer of harm or risk so it’s not just the harm of what people can see on these platforms but it also introduced the harm of the interaction with the systems especially when you think about like generative AI and recommendation algorithms, they influence our attention, our emotional state, our beliefs, the decisions we make, our behaviours,” Kamales told Scoop.
The issue was also discussed in relation to proposed measures to restrict social media access for users below the age of 16, which would require parental or guardian involvement.
Kamales said safeguarding children from harmful online material remained essential, but restrictions alone may not be sufficient to address the broader impact of AI-powered platforms and algorithm-driven content on young users.
She said online safety regulations should consider three key layers: the content users consume, how they interact with intelligent systems, and the eventual impact those interactions have on their thinking and behaviour.
“What ONSA as a next step could look at as well is beyond the content layer you need to look at the interaction layer how people are interacting with these intelligent systems and then the next layer is the cognitive layer,” she explained.
The concern is increasingly relevant with the rise of conversational AI tools and AI companions, which are designed to simulate human-like interactions and may influence users’ emotions, perceptions and choices.
Kamales said AI governance should also take into account issues such as compulsive engagement, cognitive overload and emotional dependency, while incorporating neuroscience research to better understand the effects of technology on people.
“Neuroscience can help regulators understand that human impact behind the technology rather than just regulating at a technical or process level, you can start regulating the impacts of the outputs as well on people,” she said.
However, she stressed that stronger online safety laws should not be used to control individual opinions or viewpoints.
“We don’t want to regulate what people are thinking or what opinions people have or, you know, someone might have an opposing opinion. We don’t want to regulate those opinions. We want to regulate how AI and these technology platforms influence our opinions,” she added.
She said Malaysia also needed to strengthen public awareness of how algorithms and AI-driven platforms affect users, while improving governance approaches across schools, organisations and technology platforms.
“The cognitive governance is still, there’s still a significant gap.”
“What we’re seeing is most countries are at that sort of same level where they’re looking at similar kind of approaches to governance and regulation. But the cognitive governance is still, there’s still a significant gap,” she added.
Kamales said the challenge had become more pressing as technological advances continued to outpace regulatory development.
“The technology is moving significantly faster than the governance that’s been developed. So governments are almost playing a catch-up game with the technology and this is a major issue at the moment,” she said.
Under the Online Safety Act 2025 (ONSA), Malaysia is set to restrict individuals below 16 from registering social media accounts as part of wider efforts to enhance online protection for children.
The legislation places greater responsibility on social media and online platform operators to prevent children from accessing harmful content and services, with the Malaysian Communications and Multimedia Commission (MCMC) responsible for overseeing compliance and developing suitable safeguards. – August 20, 2026