Artificial intelligence (AI) is not necessarily creating new forms of crime, but it is making existing tactics such as scams, phishing and data theft easier and cheaper to carry out, according to AI Malaysia Innovation Director Mohd Alhafidz Yahya.
He said activities that previously required organised syndicates or large call centres can now potentially be carried out by individuals using AI. Scammers can also generate more convincing and personalised messages, while voice and facial cloning have made fraudulent content increasingly difficult for victims to identify.
Alhafidz noted that AI has also improved scammers’ ability to communicate naturally in Bahasa Melayu. Messages that previously contained obvious language errors can now appear fluent and targeted. Businesses and organisations face similar risks when criminals use AI to impersonate trusted individuals and issue fraudulent instructions, including requests to transfer money.
He also stressed that AI safety controls may not perform equally across different languages. Systems extensively tested in English could have stronger safeguards than those operating in Bahasa Melayu. Malaysia’s AI Safety Institute (My AI SAFE) has therefore been tasked with testing AI models and systems, including their performance and safety controls in Malay.
Meanwhile, the government is finalising an AI Governance Bill following the completion of public consultations. The legislation is expected to be tabled around 2027, with the aim of establishing a framework for responsible AI development and use in Malaysia.
