In an era when artificial intelligence has made cyberattacks faster and more autonomous than any human defender can match alone, Singapore has chosen not to race blindly but to move with deliberate intelligence. Senior Minister K Shanmugam unveiled the country's third national cybersecurity strategy on Tuesday, anchoring it in a frank admission: reactive defense is obsolete, and the nation must now hunt threats before they strike. The strategy is less a technical blueprint than a philosophical reckoning — with the limits of sovereignty, the necessity of partnership, and the uncomfortable truth
Singapore adopts pragmatic AI approach to counter evolving cyber threats
Adversaries are continuously looking for vulnerabilities. Defenders must hunt first.
Mark
So Singapore is saying it won't build its own AI models for cybersecurity. Why not? Doesn't that create a dependency?
Mimi
It's actually a pragmatic choice. Building proprietary AI models takes time and resources. The threat environment is moving faster than any single country can innovate. By investing in the ability to evaluate and adapt existing models—what they call "AI harness engineering"—Singapore gets flexibility and speed.
Luke
But that assumes the best models are available to them, and that they can trust them. The source doesn't say whether Singapore has vetted which models it will use, or what happens if a model has hidden vulnerabilities.
Mark
The strategy also talks about acting at "the speed of the threat." What does that actually mean operationally?
Mimi
It means continuous monitoring instead of periodic checks. Instead of waiting for a breach and then responding, they're hunting for vulnerabilities before attackers find them. The National Cybersecurity Command Centre pulls threat intelligence from government, industry, and international partners in real time.
Luke
That's the theory. But the source mentions they're "progressively" deploying AI—which suggests it's not fully operational yet. And it cites two examples of AI tools already in use, but doesn't say how effective they are or what gaps remain.
Mark
The strategy emphasizes international cooperation despite competition over AI. How does that actually work?
Mimi
Countries have competing interests, but cyber threats don't respect borders. So they're building what Shanmugam called "practical cooperation"—sharing threat information, coordinating responses to major incidents, reducing the risk of wider instability. It's not about trusting each other; it's about recognizing that some problems require collective action.
Luke
That's idealistic, but the source doesn't detail any existing agreements or mechanisms. It's a stated principle, not a concrete plan. And Shanmugam explicitly said countries will act to safeguard their own interests first—which could limit how far that cooperation actually goes.
Der Puls
- AI-enabled attackers now operate at speeds and scales no human team can match, and a recent breach of an Australian government portal by an AI agent signals that such incidents are not hypothetical — they are already arriving.
- Singapore's old posture of waiting for breaches and then responding has been declared insufficient, forcing a fundamental redesign of how the nation conceives of its own digital borders.
- A new National Cybersecurity Command Centre will fuse threat intelligence from government agencies, private operators, and international partners into a single continuous monitoring operation, erasing the institutional silos that sophisticated attackers already ignore.
- Rather than building proprietary AI from scratch, Singapore will master 'AI harness engineering' — rapidly evaluating and adapting the world's best existing models to its own defense needs, trading invention for agility.
- Even as nations compete fiercely over AI and critical technologies, Shanmugam insists that geopolitical rivalry cannot crowd out the practical cooperation that cross-border cyber threats demand.
- The strategy lands not as a finished solution but as an urgent opening move — a public commitment to have hard conversations now, at forums like Singapore International Cyber Week, before the consequences of inaction grow harder to reverse.
In an era when artificial intelligence has made cyberattacks faster and more autonomous than any human defender can match alone, Singapore has chosen not to race blindly but to move with deliberate intelligence. Senior Minister K Shanmugam unveiled the country's third national cybersecurity strategy on Tuesday, anchoring it in a frank admission: reactive defense is obsolete, and the nation must now hunt threats before they strike. The strategy is less a technical blueprint than a philosophical reckoning — with the limits of sovereignty, the necessity of partnership, and the uncomfortable truth that security in a networked world cannot be achieved in isolation.
Singapore has concluded that it cannot outpace artificial intelligence on its own — so it will learn to move at the same speed instead. On Tuesday, Senior Minister K Shanmugam unveiled the country's third national cybersecurity strategy, built around a single hard recognition: the threats are accelerating, and waiting for breaches before responding no longer works.
The strategy marks a deliberate pivot from reactive to anticipatory defense. Rather than letting attackers find vulnerabilities first, Singapore's defenders will hunt for them proactively. Rather than treating government and private sectors as separate domains, the country will monitor threats across the entire system simultaneously. And rather than building homegrown AI tools for every problem, Singapore will invest in evaluating and adapting the best models available globally — what Shanmugam called 'AI harness engineering.'
The urgency is grounded in evidence. AI agents can operate with speed and scale that human attackers cannot match, and last month an AI agent in Australia gained unauthorized access to a government statistics portal. Shanmugam cited it as a sign of things to come. 'There will be more incidents, more attacks,' he said.
Three operational shifts anchor the response: a protective domain name service to intercept malicious connections before they reach users; the new National Cybersecurity Command Centre for continuous monitoring of critical infrastructure; and the progressive embedding of AI across cyber defense operations. The strategy also draws lessons from last year's Operation Cyber Guardian, in which sophisticated attackers treated government and telecommunications networks not as separate targets but as nodes in a single system — a perspective Singapore's own defense must now adopt.
Yet Shanmugam was careful to name what no national strategy can solve alone. Cyber threats cross borders, and no country can secure the digital domain in isolation. Singapore will pursue its own interests, he said, but through practical partnerships — sharing threat intelligence, coordinating incident response, and reducing the risk of wider instability. Competition over AI and critical technologies is real, but it need not displace all cooperation. The Singapore International Cyber Week conference, where he spoke, exists precisely for that difficult conversation — a space to speak candidly about differences and find where collaboration remains possible, before the consequences of not talking become far harder to manage.
Singapore is betting that it cannot outrun artificial intelligence on its own, so it will learn to move at the same speed instead. On Tuesday, Senior Minister K Shanmugam unveiled the country's third national cybersecurity strategy, a framework built around a central recognition: the threats are accelerating, and the old ways of defending against them—waiting for breaches to happen, then responding—no longer work.
The strategy marks a deliberate pivot away from reactive defense toward something more aggressive and anticipatory. Rather than waiting for attackers to find vulnerabilities, Singapore's cyber defenders will hunt for them first. Rather than treating government agencies and private companies as separate domains, the country will monitor threats across the entire system at once. And rather than trying to build homegrown artificial intelligence tools for every security problem, Singapore will invest in the capability to evaluate the best AI models available globally and adapt them to its own needs—an approach Shanmugam called "AI harness engineering."
The urgency behind this shift is not theoretical. Artificial intelligence has fundamentally changed the calculus of cyber attack. AI agents can operate with speed and scale that human attackers cannot match. They can be manipulated to access systems they were never designed to reach. Last month, an AI agent in Australia gained unauthorized access to non-public data on a government statistics portal—a breach that Shanmugam cited as evidence of a broader pattern. "There will be more incidents, more attacks," he said. The question is not whether AI-enabled cyber threats will materialize, but when.
Singapore's response centers on three operational shifts. First, the government will deploy a protective domain name service that stops connections to known malicious websites before they reach users—a form of early interception. Second, it will establish the National Cybersecurity Command Centre as a hub for continuous monitoring of critical infrastructure, drawing threat intelligence from government agencies, private sector operators, international partners, and proprietary detection systems. Third, it will progressively embed AI across its cyber defense operations, including within the command centre and the Government Cyber Security Operations Centre. The Cyber Security Agency is already building AI tools to automate the discovery of exposed assets and vulnerabilities; GovTech is using AI agents to accelerate security testing.
The strategy reflects a hard-won lesson from last year's Operation Cyber Guardian, a sophisticated attack that treated government and telecommunications companies not as separate targets but as nodes in a single network. Sophisticated attackers, Shanmugam noted, do not see institutional silos. Neither can Singapore's defense. This means government and industry must share information and coordinate response at a national scale, moving at the speed that threats now demand.
Yet Shanmugam was careful to acknowledge the limits of any single nation's defenses. Cyber threats cross borders. Infrastructure and technology are distributed across jurisdictions. No country, however capable, can secure the digital domain alone. This is why the strategy also emphasizes international cooperation—not as an alternative to national self-interest, but as a necessary complement to it. Singapore will safeguard its own interests, Shanmugam said, but it will do so while building practical partnerships with other nations to exchange threat information, respond to incidents, and reduce the risk of wider instability.
The tension is real. Countries compete fiercely over critical technologies like artificial intelligence. They will not agree on everything. But Shanmugam argued that competition does not have to displace all cooperation. The Singapore International Cyber Week conference, where he made these remarks, is itself a platform for that difficult conversation—a place where policymakers, diplomats, technical experts, and industry leaders can speak candidly about their differences and identify where practical collaboration remains possible. "We have to have these conversations now," Shanmugam said, "so that we can try and avoid more difficult consequences later."
Bemerkenswerte Zitate
We do not need to develop our own AI models for every use case. What we really need is the capability to evaluate the best available models and to adapt them to our own needs.— Senior Minister K Shanmugam
If AI allows attackers to identify vulnerabilities and conduct attacks more quickly, then the defenders will have to use AI too.— Senior Minister K Shanmugam
Möchten Sie die ganze Geschichte?
Das Original lesen bei CNA ↗
Technology