On October 3, Chunghwa Telecom reported a fault in the Trans-Pacific Express submarine cable and said it had rerouted traffic through other international cables. While repairs were underway, connections to China during peak hours could experience slight delays. By October 23, Taiwan's Ministry of Digital Affairs still listed the fault, with repairs estimated for November 30. Neither the operator's notice nor the ministry's status table attributed the fault to deliberate interference. Communication continuity is an integral part of Taiwan’s defense strategy. For a country where communication disruption is a considerable concern, this raises a broader question: how should Taiwan respond to a communications failure before its cause has been established?
I would argue that Taiwan should prepare to act on that uncertainty. Restoring communications should not require the same level of evidence as accusing Beijing of causing their failure. The island can move traffic, protect essential services, and tell the public what it knows while investigators work to establish responsibility. Although this may sound like a modest distinction, in a crisis, tying those decisions together would give whoever caused the disruption considerable influence over how quickly Taiwan could respond. A damaged connection should trigger a recovery plan but not require an immediately identified culprit.
The evidence from Taiwan's own government makes caution necessary. In its April review of cable damage, the Ministry of Digital Affairs identified ship anchors as a leading cause of nearshore incidents and earthquake-triggered submarine landslides as the main offshore risk. The report also described criminal convictions in two 2025 cable-damage cases. Deliberate or unlawful interference deserves serious attention. So does the possibility that an incident has a different cause. Recognizing both is essential if Taiwan wants its accusations to carry weight when the evidence does support them.
There are several questions that need answering before a broken cable can be called a state-directed attack. What physically failed? Which activity caused the damage? Was it deliberate? Was someone acting under government instructions? Evidence can answer one question while leaving the others unanswered. Even identifying a vessel responsible for damage does not, by itself, establish who ordered it. This matters because policy often moves faster than the investigation. A government may have to authorize emergency assistance while it is still deciding whether it has grounds for a public accusation.
The information environment makes that separation especially important this month. On October 21, Reuters reported that Taiwan's National Security Bureau had identified more than 2.5 million items of China-linked disputed or false online information during the first nine months of 2026, roughly 48 percent more than during the same period in 2025. The bureau warned of interference ahead of November's local elections. Beijing has previously denied election meddling and disinformation allegations. These figures describe the bureau's assessment of detected content; they do not measure how many people changed their minds or establish any connection to the October cable fault.
They do, however, help explain why the public explanation of an outage deserves attention alongside its repair. In a January study, researcher Aurelio Insisa examined how cable disruptions can spill into the information environment and weaken confidence in Taiwanese institutions. An implication that can be drawn from this is that Beijing would not have to cause every failure to benefit from confusion surrounding one.
Taipei should not assume that public anger over communications failures is evidence of a Chinese influence campaign. Someone who cannot make a call or process a payment has a legitimate complaint. Dismissing it would deepen the very mistrust officials hope to prevent. A credible response needs to explain the service problem, acknowledge the inconvenience, and distinguish verified information from an assessment that may change. Officials cannot reasonably ask the public to accept uncertainty while presenting their own assumptions as settled facts.
Taiwan is already investing in tools that could make the investigation faster. When the government announced the opening of the Taiwan-Matsu No. 4 cable in September, it also described a warning system combining AI recognition with vessel-identification data. The new cable adds capacity and works alongside microwave and satellite connections to improve backup options. These are practical steps, but the remaining question is how an alert generated by a monitoring system becomes a judgment that an official is prepared to defend.
A system that flags unusual activity can help direct scarce attention. Its alert, however, still needs an explanation. An unexpected vessel movement might warrant a closer look without proving sabotage. Conversely, an activity that fails to trigger an alert may still deserve investigation. The value of the system depends partly on whether investigators can examine what it detected, what it missed, and what other information supports their conclusion.
“AI can accelerate the extraction of useful insights, but it does not eliminate the need for manual research,” Stiepan A. Kovac, CEO of Quantum Resistant Cryptography, said in an email exchange with Strait Perspective. “In ambiguous environments, the analyst still has to distinguish between false positives, false negatives, and what the available evidence actually supports.”
That is a useful standard, provided that manual review means more than having someone sign off on a machine's answer. People bring expectations to an investigation, too. NIST's AI Risk Management Framework warns that human interaction with AI can, under some conditions, amplify bias. For Taiwan, the practical question should be whether an analyst has access to corroborating evidence and the authority to challenge a result. If the institutional pressure is to confirm the first alarming explanation, adding a human reviewer will accomplish very little.
The response to a communications failure should therefore begin with its consequences. Is traffic being rerouted successfully? Are hospitals, emergency services, and local governments able to communicate? Which communities face a prolonged interruption? Those questions should determine the initial protective measures. Chunghwa Telecom's October notice already demonstrated the basic approach: it described the problem, explained that traffic had been moved, and identified the inconvenience customers might experience. The notice did not need to settle responsibility to give people useful information.
There are also limits to what backup infrastructure can accomplish. A May report by Jacob Stokes and Ryan Claffey at the Center for a New American Security explains that satellite connections can supplement Taiwan's submarine cables but cannot replace their overall capacity. Terminals and ground infrastructure also matter. Buying access to a satellite service does not automatically put a working connection where it is needed. In a severe disruption, officials may have to prioritize limited capacity. Those choices belong in exercises and service agreements before an emergency arises. This way they can be properly debated without a damaged network influencing the deadline.
Public communication needs the same preparation. A publicly shared incident update that distinguishes the observed failure from the suspected cause and explains the confidence behind any attribution could help preserve public trust. An update that illustrates to the general population what services remain available and when they can expect the next update would effectively build on Insisa's call for coordinated communication among the agencies handling cable incidents. A promise to provide another update is useful even when investigators do not yet have a culprit to name. Silence, on the other hand, leaves everyone else to supply their own explanation.
A worthwhile exercise would test what happens when the first explanation turns out to be wrong. Officials could begin with a suspected attack and then receive evidence of accidental damage, or start with an apparent equipment fault and later discover deliberate interference. Could the response adjust without losing time or concealing the correction? Would essential services keep operating throughout? An exercise that rewards a quick, confident accusation may train people for the wrong task. Taiwan needs decisions that remain useful as the evidence changes.
Its partners should apply the same distinction. Assistance with repair coordination, emergency connectivity, and technical investigation should not depend on an immediate agreement about Beijing's involvement. Such support can be justified by the disruption itself. That would also make assistance harder to derail through disputes over an incident's origin. The United States and other partners can help Taiwan retain its ability to function while reserving a separate decision about sanctions or other consequences for whoever is responsible.
There is a serious objection to consider here. If Taiwan concentrates on recovering from disruption, does it risk teaching an adversary that interference carries no punishment? It would, if restoration became a substitute for investigation. That is, however, not the approach that is illustrated here. Evidence should be preserved as recovery proceeds, and credible findings should lead to proportionate consequences. Public attribution should explain its basis as far as security permits, without requiring officials to disclose every intelligence source. The distinction is between acting before every question is answered and pretending that the unanswered questions no longer matter.
This also changes what should count as a successful investment in defense. A warning system that produces more alerts is not necessarily making Taiwan safer. Officials should ask whether it helps them identify consequential incidents and restore essential connections. Spending on repair access and rehearsed procedures may attract less attention than a new platform, yet these are the things that determine whether equipment remains useful when several failures arrive at once. Taiwan should expect its resilience programs to demonstrate those results, rather than treating the purchase itself as the achievement.
The October cable notice offers a restrained example of how to begin. It identified a fault and explained what was being done about it. A more severe incident would demand much more, but it would not remove the need for that discipline. Taiwan has good reason to investigate interference aggressively. It also has reason to protect its credibility when the evidence is incomplete. Its people should not have to wait for a definitive answer about responsibility before their government can keep them connected.