The Pacific Drone Arms Race Has No Rules

Abstract

This article examines how the rapid spread of aerial, surface, and subsurface uncrewed systems across the Pacific is creating new risks of miscalculation, particularly around Taiwan. It argues that existing mechanisms such as the Code for Unplanned Encounters at Sea (CUES) are poorly suited to autonomous and remotely operated platforms, leaving unresolved questions of identification, attribution, human control, and escalation. The article proposes supplementary regional encounter protocols to reduce the risk that an ambiguous drone incident becomes the opening move of a wider conflict. 

Introduction

On June 16, a small autonomous surface vessel operated by the California-based company Seasats moved close enough to a Chinese Type 052D guided-missile destroyer to record detailed footage of it. The encounter took place about 65 miles northwest of the Philippine Island of Luzon, inside the Philippines’ exclusive economic zone. No one was injured, no warning shots were fired, and no sailors faced off against each other. But the incident previewed a dilemma regional governments have just started to face: How to respond to an unmanned boat of unknown intent when it approaches a warship in a contested area?

Ferdinand Magellan gave the Pacific Ocean its name during his expedition in 1520. Five hundred years later, the world’s largest ocean is seeing a storm brought on by military systems that carry neither pilots nor sailors. This change is more than just naval rivalry. The East Asia and Pacific region is home to 32% of global trade and has become the hub of global maritime activity. A crisis in the Pacific would thus not only threaten the forces themselves, but also ports, shipping routes, supply lines, and economies well beyond the ocean.

The Pacific’s Drone Buildup

The U.S., China, Australia, Japan, Taiwan, and others in the region are all developing their uncrewed capabilities. Washington’s Replicator program aims to deploy many small, expendable, autonomous systems, and the US Pacific Fleet has created a squadron of small uncrewed surface vessels. China has been showing increasingly advanced aerial, surface and underwater systems and has been using drones in exercises around Taiwan. Australia’s MQ-28 Ghost Bat flew in formation with a US F-15EX for Exercise Valiant Shield and conducted dispersed refueling from Rota in the Northern Mariana Islands. Japan is developing a defense system that combines uncrewed and manned aircraft, and Taiwan is increasing domestic drone manufacturing as part of its asymmetric defense policy.

There are good reasons for this rush. Such systems can conduct surveillance, resupply isolated units, search for submarines, relay communications and expose hostile positions without risking the lives of human crews. They can be deployed in certain missions at lower cost than conventional aircraft or warships and can also be deployed in missions that are not possible with conventional equipment. Drones would make a Chinese blockade or amphibious assault of Taiwan more difficult. For China, they can expand surveillance, targeting, and pressure around the island. Lastly, for the U.S. and its allies, they offer means to spread military power out over a large maritime area.

When Expendability Creates Risk

This should make crises easier to contain at first glance. When an uncrewed aircraft or vessel is destroyed, no pilot or sailor is killed, providing governments with more leeway to take the loss without retaliating. But the same expendability can encourage risk-taking. States and even private operators could be more inclined to fly drones near enemy forces, contested areas, military bases, and sensitive shipping lanes, since there’s no one on the ground in danger. For a crewed platform, a mission that might otherwise be too provocative to authorize could be more easily authorized when no one is aboard.

The real issue is uncertainty. An officer observing an oncoming uncrewed vessel may not be able to determine if it is taking photographs, gathering electronic intelligence, carrying a jammer, placing a sensor, or preparing an attack. An aircraft flying towards a warship could be remotely controlled on a pre-programmed course, under autonomous guidance, or simply be faulty. There are few clues about intention and fewer opportunities for direct communication before a decision needs to be made if there is no crew on board. In an even more harrowing scenario, an uncrewed system can misidentify a civilian settlement, vessel, or aircraft for a military target and act against it, leading to civilian loss of life and property and necessitating a strong military response from the nations involved.

Taiwan and the Risk of Miscalculation

Such danger and uncertainty could become especially dangerous around Taiwan. Chinese military aircraft and vessels are already present in the region, as are Taiwanese forces, heavy civilian maritime and air traffic, and occasional American and allied operations. China’s late-2025 Taiwan exercises featured ships, aircraft, missiles and drones, including scenarios of blockade and attacks on maritime targets. Meanwhile, Taiwan is also developing a “hornet’s nest” of air, surface and subsurface drones that could help fend off an invasion, according to American officials.

The next crisis could then start with a sense of uncertainty as the following situation demonstrates. A reconnaissance drone enters a sensitive area during a major exercise. The other side jams it, captures it, or shoots it down. Other unmanned systems are dispatched to find it; ships start to move; planes take off, and both sides claim the other is escalating. The situation can be dangerous without anyone wanting a war. Leaders would still have to make quick decisions without knowing what the platform is doing, who is controlling the platform, and whether its behavior is intentional.

Conflict is not inevitable, and it is possible these systems deter deliberate attacks. However, they will open up more avenues for surveillance, probing, interference and misunderstanding. The main threat is not that a machine will make its own decision to go to war. It is that uncrewed systems expand the number of encounters in which commanders must interpret intent without the usual human signals.

The Rules Have Not Kept Pace

Current regional regimes only partially address this issue. In 2014, navies in the region endorsed the Code for Unplanned Encounters at Sea. It offers voluntary means of communication and safe maneuvering during naval ship and aircraft encounters. Unfortunately, the non-binding agreement predates the current proliferation of autonomous military systems, and doesn’t offer any specifics on how to identify the operator of an increased platform, how to contact the remote controller, or how to respond if the platform ignores warnings.

Civilian control is progressing, but it does not solve the military issue. The International Maritime Organization (IMO) approved its first global safety code for autonomous vessels in May 2026. The code covers navigation, remote operations, and cybersecurity, but is not mandatory and mostly for commercial shipping. It was not intended to regulate interactions between military drones and foreign warships in sensitive political waters.

The Seasats incident off Luzon highlights another complicating factor: private ownership. A warship may not be aware of whether the drone it encounters is being used for a commercial client, to test technology, under a government contract, or as part of an intelligence mission. The platform might have little or no national markings, and the people operating it might be hundreds of kilometers away. The distinction between private and state activity will become harder to discern in a crisis as militaries increasingly depend on commercially developed systems.

Regional governments need to start negotiating a supplementary code for uncrewed military encounters before a serious incident forces the issue. Such a framework need not settle sovereignty disputes or restrict lawful military operations. It should set up recognizable identification, verified contact information for a responsible human operator, minimum separation distances, common warning procedures and a clear escalation procedure before the platform is jammed, captured or destroyed. It should also make clear that states remain responsible for privately owned systems conducting military or security-related missions.

No major power will disclose the full capabilities, payloads or autonomy software of their systems, nor should an agreement demand that. Rules should be about behavior that can be seen. A drone doesn’t have to reveal all of its sensors. Still, the operator should be held responsible for reacting to warnings, keeping safe distances and avoiding actions that another military might reasonably interpret as an imminent attack.

Australia and Japan are in a good position to champion such an initiative. Both heavily invest in uncrewed systems, have close ties with the United States, and are members of regional forums that involve China and Southeast Asian countries. Australia played a role in the development of CUES, and Japan’s growing drone program means it has a vested interest in ensuring that autonomous platforms do not cause uncontrolled escalation. Regional or bilateral agreements may follow initial agreement under the Western Pacific Naval Symposium.

Such rules will not eliminate the climate of mistrust in the South China Sea, nor will they solve any sovereignty issues or stop any intentional coercion. They are more specific in their role: to minimize the risk of a tactical engagement with an uncrewed platform being confused with the first phase of a larger attack.