AI News: Malaysia plans its first AI law, expected to be tabled in early 2027
On 5 October 2026 Digital Minister Gobind Singh Deo told the Dewan Rakyat that an AI Governance Bill is being drafted with a risk based approach. Here is what has been announced, and what has not.
Malaysia is drafting its first law dedicated to artificial intelligence (AI). On 5 October 2026, Digital Minister Gobind Singh Deo told the Dewan Rakyat that the Artificial Intelligence Governance Bill (RUU Tadbir Urus Kecerdasan Buatan) is expected to be tabled in early 2027.
He was answering an oral question from the Kubang Pasu MP, Datuk Dr Ku Abd Rahman Ku Ismail, on the government's position on global calls to slow AI development. The Gombak MP, Dato' Seri Amirudin Shari, also asked about AI development.
What has been announced
- Risk based. According to Bernama, the level of controls and obligations will be assessed by the potential harm of the AI system itself, not by industry alone. Uses that directly affect safety, fundamental rights and individual wellbeing will face stricter oversight.
- Cross sector. The bill is designed as a framework across all sectors, to complement existing sector laws rather than replace them.
- Safety before wide use. For increasingly powerful and autonomous AI systems, Gobind said safety controls, testing, monitoring, human oversight and response mechanisms must be strengthened before wider deployment. "Priority will always be given to safety," he said, as quoted by Bernama.
- Malaysia AI Safety Institute. The government has set up the Malaysia AI Safety Institute (MY-AISafe) to study clearer ways to regulate how organisations use AI models. Its responsibilities include safety benchmarks, risk assessment and incident reporting so corrective action can be taken.
According to Media Selangor, the move is in line with the National AI Action Plan 2026 to 2030 under the "Negara AI 2030" direction.
What we do not know yet
The bill is still being drafted. As of the 5 October reports, several key details have not been disclosed:
- Which AI uses count as "high risk"
- Penalties for breaking the law
- Which agency will enforce it
- The exact tabling and commencement dates
So be careful with social media posts claiming to know specific fines or bans. The bill's text has not been published.
What it means for you
If you are an everyday user of ChatGPT, Gemini or other AI tools for daily work, nothing changes today. Risk based approaches usually focus on systems that make important decisions about people, not everyday uses like writing or summarising.
If you are a creator or small business owner using AI for content, ads or customer service, this is a good time to build habits that may be asked of you later:
- Keep a record of which AI tools you use and for what.
- Review AI output before publishing, especially facts, prices and claims.
- Do not put customers' personal data into AI tools without a clear reason.
- Be honest with customers when they are chatting with a chatbot, not a person.
If your business builds or sells AI systems that assess people, for example for hiring, loan approval or monitoring, follow this bill closely. Those are the categories most tightly controlled in risk based AI laws elsewhere.
We will update this article when the bill's text is published. Meanwhile, the AI tools we use are listed on the AI Tools page.
Sources
- RUU Tadbir Urus AI Dibentang Di Dewan Rakyat Awal 2027: Gobind (Bernama, 5 October 2026, in Malay)
- RUU Tadbir Urus AI dijangka dibentang awal 2027: Menteri (Media Selangor, in Malay)