Singapore launched its third cybersecurity strategy on 13 October, setting out plans to use AI models selected from outside government in national cyber defence rather than develop a model for every task. Senior Minister and Coordinating Minister for National Security K Shanmugam announced the strategy at Singapore International Cyber Week, CNA reported. The government plans to introduce AI progressively into the National Cybersecurity Command Centre (NC3) and the Government Cyber Security Operations Centre.
Key points
- The strategy is planned for implementation over five years.
- Singapore will invest in evaluating, adapting and securely integrating AI models for cyber defenders.
- NC3 will draw together threat information from government, critical infrastructure owners, companies and international partners.
- A planned protective domain name service will block connections to known malicious domains under the Online Criminal Harms Act.
Five years for the Cybersecurity Strategy 2026
The Cybersecurity Strategy 2026 has four strategic pillars and two foundational enablers, with implementation planned over five years, TechEdt reported. Its priorities extend from critical infrastructure and the wider digital environment to AI security, quantum-safe migration and international cooperation. The earlier strategies were launched in 2016 and 2021.
Shanmugam said Singapore would invest in “AI harness engineering”: the ability to assess available models, adapt them to local needs and connect them securely to the data, tools and workflows used by cyber defenders. “We do not need to develop our own AI models for every use case,” he said. That approach puts model selection and integration inside the government’s defence operation while leaving room to choose among models as the technology changes.
Shanmugam argued that defenders need to search continuously for weaknesses and hostile activity, rather than rely on periodic assessments and responses after an incident. He said attackers could use AI to operate faster and at greater scale. As an example of the risks associated with AI agents, he cited an agent that gained unauthorised access to non-public data on an Australian government statistics portal in September. AI Affairs has previously reported that Singapore was studying stronger safeguards for high-risk AI uses.
NC3 will coordinate national cyber defence
The Cyber Security Agency of Singapore (CSA) is enhancing NC3 as its national cyber defence operations hub. Shanmugam said the centre would combine information from government, owners of critical information infrastructure, cybersecurity companies and international partners to track vulnerabilities and threats. It is intended to support continuous monitoring of critical infrastructure and help defenders detect persistent attackers earlier.
CSA is developing AI tools to map exposed assets and vulnerabilities and assist specialists testing weaknesses in web applications. GovTech uses AI agents to speed up vulnerability discovery and security testing, Shanmugam said. The government also plans to work with industry and technology partners to extend AI-enabled defence capabilities across critical sectors. Within NC3, AI-enabled coordination systems are intended to help organise a national response to major threats.
The strategy draws on Operation Cyber Guardian, announced in February 2026. More than 100 cyber defenders from government agencies worked with telecommunications operators in the response to an attack by the advanced persistent threat actor UNC3886 against Singapore’s telecommunications sector. The operation was described as Singapore’s largest coordinated cyber incident response, TechEdt reported.
CSA and the Centre for Strategic Infocomm Technologies are progressively equipping selected critical infrastructure owners with proprietary detection systems and integrating them with NC3. The strategy also calls for monitoring exposed internet-facing assets and working with agencies and infrastructure owners to address potential entry points. Those measures give the national hub a role in finding risks across operators as well as coordinating action during an incident.
Domain blocking under the Online Criminal Harms Act
The government plans to develop a national protective domain name service that can stop a connection to a known malicious domain before it reaches a user. Shanmugam presented it as a way to reduce exposure to phishing, scams and malware. The service is intended to operate under the Online Criminal Harms Act, with assessment and validation before a domain is blocked to reduce incorrect decisions, according to TechEdt.
The strategy also addresses the security of AI systems adopted by organisations. The government intends to prepare implementation guidance for secure adoption, while working with public agencies and industry to accelerate migration towards quantum-safe systems. People and entities whose access to a site is restricted may request a review of the decision or challenge it through established channels.
