Get the People Off the Boat: Unmanned Is the Essence of Autonomy

Although the machine gun had the potential to concentrate the firepower of a rifle company in the hands of a small group of infantry, not everyone saw things that way. During the Franco-Prussian War, France deployed its mitrailleuse, its first machine gun, not with infantry formations, but with artillery batteries located behind the front lines. Firing at long range, the artillerymen had difficulty identifying where their plunging fire was striking, causing delays in adjusting their aim, and thus rendering the mitrailleuse largely ineffective. The soldiers who were subjected to its fire were especially dismissive of the weapon.[1]

In hindsight, this misuse of the mitrailleuse appears inexplicable, but as historian and strategist Colin Gray was fond of reminding us, combat itself is the only true test of new technology, weapons, and doctrine.[2] A weapon’s impact can vary from engagement to engagement because circumstances, and the opponent’s idiosyncrasies, play a part in shaping its effectiveness.[3] It is not always entirely clear how a new weapon should be deployed and used before the test of battle.

Given an ever-expanding array of technologies, the US Navy faces many decisions about which technologies to weaponise and how best to use new weapons. The most important decisions concern how best to develop and employ autonomous systems, especially systems that can substitute for crewed surface ships and submarines. Put somewhat differently, if the machine gun’s essence makes it a front-line infantry weapon, what is the essence of the autonomous surface or subsurface vessel? How can artificial intelligence best be integrated into autonomous systems and platforms to increase the combat effectiveness of the Navy? How should autonomous systems be designed, deployed, and used to maximise their contributions?

As the USN moves ahead with autonomous vessels, it will have to grapple with the use of AI, command-and-control, and manning concepts. Here, an AI-generated image shows future unmanned surface vessels operating in formation. SORA.

The Drone Revolution

Judging from the headlines, it would appear that a revolution in warfare is underway, and drones are at the heart of that revolution. Semiautonomous weapons in the form of mines, cruise missiles, point-defense systems, and remotely piloted aircraft have been a fixture in arsenals for decades, but the availability of ever-increasing amounts of computational power in ever-smaller packages is creating what could be a new era of autonomous warfare. For instance, small and inexpensive aerial drones dominate the battlefields of Ukraine.[4] In the Black Sea, semiautonomous skiffs packed with high explosives pose a lethal threat to Russian surface combatants, forcing them to withdraw out of range.[5]

In a spectacular demonstration of the potential of drones, Ukraine recently smuggled aerial vehicles deep inside Russia to destroy long-range bombers that were being used to launch cruise missiles against urban targets (Operation SPIDERWEB).[6] With only slightly less fanfare, Israeli Special Forces launched drones from inside Iran to degrade air defenses at the outset of the air campaign against Iran’s nuclear program.[7] While the success of these operations was based as much on ingenuity and daring as the drones themselves, there is no doubt semiautonomous weapons are now a fixture of the battlefield. It is only a matter of time before fully autonomous weapons are available.

The French mitrailleuse early machine gun was misused during its combat debut. Shutterstock.

The problem the USN faces, especially as it contemplates operations in the vast expanse of the western Pacific, is that its autonomous vessels are not derived from commercial systems built for hobbyists or limited commercial applications. Many drone vessels under construction for the Navy are large, with displacements in the hundreds of tons. Even the least inexpensive Navy models cost millions of dollars. These vessels sometimes look like oddly-designed ships because they can be optimised without the sea-keeping qualities necessary to keep human operators comfortable. Nevertheless, many of these platforms resemble commercial vessels used in the petroleum industry or specialised ships used to carry commercial cargo containers because they are based on existing crewed vessel designs. They have plenty of open deck space to carry mission containers that give them the ability to act as decoys; undertake intelligence, surveillance, and reconnaissance missions; or even act as ‘missile barges’ to bolster the offensive firepower of crewed warships that constitute the bulk of the fleet.

In truth, the larger ‘autonomous’ vessels that will soon be delivered to navies everywhere are based on existing designs for crewed ships. As a result, they could carry small crews to operate the platform some or even all of the time. Designers and strategists are considering various approaches to placing human operators on the bridge of these ‘autonomous’ and ‘semiautonomus’ vessels. This is akin to hanging a lawn chair under a quadcopter to provide a human operator with a bird’s eye view of the battlefield.

Technology: Essence and Ideology

Philosophers of science have advanced the idea that ideology is embedded in technology and gives it an essence that shapes the world around it and the way people use it.[8] Martin Heidegger, for instance, noted how technology reorders the way people view the world as something that can be consumed by or manipulated by that new technology.[9] With the advent of steam-powered pumps that enabled deep mining, the view of the countryside changed: “The earth now reveals itself as a coal mining district, the soil as a mineral deposit.”[10] Heidegger was suggesting that there is a law of the instrument – give a man a hammer and he will discover everything needs pounding.

By contrast, cultural researcher Robert Kozinets suggests technology is embedded within a broader ideology that shapes a narrative guiding its adoption and use. There are various ideologies, but the incorporation of artificial intelligence (AI) into maritime platforms seems to reflect techtopianism – the idea that new technology inevitably improves human affairs, despite twists and turns along the way.[11] Techtopianism, however, has a downside. It can lead designers and operators to assume only benefits of autonomy and that those benefits are easily achieved without much thought to weaponisation or integration into force structure, doctrine, and national strategy. Techtopianism can obfuscate the essence of a particular technology because it can lead developers to take a lackadaisical attitude toward its application (weaponisation).

History suggests technology badly weaponised or poorly employed will not contribute to combat success and might detract from a fleet’s overall effectiveness. The weaponisation and employment of new technology must align with its inherent essence if its qualities are to manifest in battle.

Assessing autonomous maritime systems is tricky because no list of characteristics best describes their essence. Even if there were, it would always be possible to identify other factors that shape their performance in significant ways, and these new characteristics will only be revealed by battle itself. Nevertheless, by considering five characteristics – service life, specialization, cost, lethality, and human occupancy – it is possible to identify the essence of autonomous systems and how they might be designed to bolster the capabilities of crewed vessels. In other words, what characteristics, revealed by the successful weaponisation of technology, enable autonomous ships and submarines to best contribute to the fleet?

The Orca XLUUV is a good example of pursuing autonomy appropriately. It was not built with optional manning, and thus embraced the essence of unmanned systems. US Navy.

Expected Service Life

Surface ships and submarines can last decades, limited mostly by the willingness of their operators to tolerate ever-increasing maintenance and modernisation costs. The keel of the USS Nimitz (CVN-68) was laid down in 1968. It is currently scheduled to begin (nuclear) deactivation in 2027, a relatively short 52 years after commissioning. The Hemminger (DE-746), a Cannon-class destroyer escort currently serving with the Thai Navy, was commissioned 81 years ago. Warships represent a significant capital expenditure, but they can remain in service across generations.

By contrast, the expected service life of autonomous platforms should probably be measured in months or years, not decades. Longevity requires heavier construction, redundant engineering, persistent and growing maintenance, and crews to repair equipment on board while underway – characteristics that increase cost, complexity, size, and construction time. Autonomous systems only need a service life sufficient to achieve a specific mission over a limited time before they are rendered obsolete by new technologies becoming available at decreasing intervals.[12] Building autonomous systems ‘to last’ is an impulse powered by the high cost of crewed naval warships, service culture, and tradition; it does not align with the essence of autonomy. Autonomous vessels should be built to fill a niche and last perhaps five to seven years, with an expectation that new technologies will be incorporated into a future design to meet future requirements.

Multipurpose or Specialised 

The search for economies of scale as ship production declined and costs soared, has driven USN planners to include as many mission capabilities in a single large hull as possible.[13] For example, one Arleigh Burke–class destroyer can engage in a surface action in the morning and then be reassigned to provide ballistic-missile defense in the afternoon. If two less-expensive, but mission-constrained, surface combatants – one specialising in surface warfare and the other equipped to provide air and missile defense – were deployed to cover these missions, their total construction, manning, operating, and logistics costs could exceed significantly the construction, operational, and personnel costs of a single Arleigh Burke destroyer.[14] As the number of platforms constructed declined, the number of their organic capabilities increased.

Surface and subsurface autonomous systems, however, should be built in far larger numbers than crewed vessels, opening the door for platform specialisation. An autonomous vessel built for air defense should be optimised for that mission alone. Squeezing antisubmarine or antisurface warfare systems into the platform would detract from its primary mission, create complexity, and increase construction time and cost. Drones should be optimized for a specific mission and not built with excess capacity to include ‘plug-and-play’ or modular systems – that is, the ability to undertake different missions as different packages are loaded on the vessel. Modular mission systems are a costly step away from mission specialisation. The fact that the USN considers modularity as something desirable in its emergent autonomous surface fleet suggests cost considerations loom.[15]

Moreover, if the goal is to use a generic box to hide various mission packages from observation, there are other ways an adversary can discover a platform’s mission, making it imperative to control various electronic signatures, operations, and lifecycle events to bolster operational security. It would be better to spend scarce resources to build more specialised systems than modular capabilities that find their way onto platforms to justify their exorbitant costs. The Navy already consists of relatively few multipurpose warships; building a few, modular autonomous vessels would not add much to overall capability.

The Ukrainian Magura V unmanned surface vessel has been a key ingredient in Ukraine’s sea denial campaign against the Russian Black Sea Fleet. The United States needs USVs like the Magura V, and larger. Ukrainian SBU.

Expensive vs Expendable

The commanders and crews of USN destroyers that supported the 6 June 1944 landings on Omaha Beach engaged shore batteries at point-blank range in water so shallow many touched bottom, risking grounding that would have made them easy targets.[16] Their exploits demonstrate the great risks commanders took, but such risk-taking is rare in today’s Navy.[17] There are so few ships today that none could be considered expendable; each represents a key part of national defense that would take years and billions of dollars to replace.

Because they should be relatively inexpensive, autonomous surface and subsurface vessels ought to be purchased in large numbers. If autonomous capability is not concentrated in a few hulls, the loss of any one could be absorbed in battle without jeopardising the entire evolution. Mission optimisation, a short service life, and ease of production should lead engineers to design systems to be expendable from the outset. Sonobuoys, for example, are autonomous systems that listen, transmit, and die. All but the smallest autonomous systems will probably cost more than sonobuoys, but the availability of a large inventory of platforms and the speed of replacement should give commanders operational and tactical options they currently lack when faced with risking exquisite, manned platforms. Once an autonomous missile barge goes Winchester, commanders can send it to the rear for rearming – or forward as a decoy, an option not available with a destroyer.

Lethal or Survivable

How should designers expend their marginal dollar – on system lethality or survivability? Autonomous systems require enough survivability to accomplish their mission, for some minimum amount of time, but given the need to keep their costs low, resources should be prioritised to increasing their lethality or system performance in combat. Engineering redundancy, hardening operational systems, or damage-control systems make little sense for platforms designed to be ‘used up’ in combat. By contrast, autonomous surface and underwater platforms – operating either singularly or as swarms – need the capability to inflict a mission kill on an enemy major combatant or to complete some other primary mission.

Strategic and operational considerations also favor lethality over survivability. In terms of strategy, autonomous vessels are a relatively low-cost way to compensate for the superior numbers an opponent possesses. The historical precedent is the creation of Howell Automobile torpedoes and torpedo boats, which were intended to threaten the supremacy of larger navies.[18] Although a torpedo boat had no chance of surviving an accurate salvo from a large warship, the torpedo gave an inexpensive and simply constructed vessel the capability to sink a major combatant, encapsulating the idea that the key to success in naval combat is to fire effectively first. Now, autonomy allows commanders to risk platforms to achieve hits. Given the asymmetric nature of the contest between a capital ship and a torpedo boat, or a crewed multipurpose warship and a drone, large numbers of expendable individual platforms provide a significant lethality advantage.

Crewed vs Uncrewed

Today, the USN is struggling with a host of operational and force-design questions related to the introduction of autonomy. According to a recent report from the 2025 meeting of the Surface Navy Association: “The service still seems to be grappling with a holistic concept of operations: when the ships will be unmanned versus optionally manned, how they’ll work alongside manned vessels, where they can assume some of the burden of manned ships, and what employment might look like in combat, among other things.”[19] All of this needs to be worked out, and quickly. But one mistake the USN must avoid in designing autonomous vessels is to house a crew continuously or optionally because that would not be in keeping with their essence. Autonomy is the key characteristic that distinguishes drone ships from today’s surface and subsurface fleet.

If the Navy decides to put people on these new platforms, it will lose the qualities most needed to make autonomous vessels effective. Construction costs will increase because a vessel capable of carrying humans will have to be more complex, survivable, maintainable, and capable. To justify the additional cost of putting humans on board, the Navy is suggesting a one-size-fits-all ‘containerised’ mission approach, but human habitability will only increase construction, operational, and maintenance costs.[20] For instance, a vessel capable of carrying different mission ‘boxes’ would have to possess the communications, power, and sea-keeping ability to carry out the mission (missile firing, ISR, decoy, etc.) contained in its box. Increased cost reduces the number of platforms that can be constructed, leading to a reluctance to treat the vessels as expendable. Moreover, it is hard to envision how crewed autonomous vessels can be constructed cheaply and quickly, reducing the prospect of designing and constructing new unmanned surface or underwater vehicles to meet emergent threats or take advantage of new opportunities.

The essence of autonomous surface and subsurface vessels is their autonomy. Autonomy creates opportunities to build large numbers of relatively inexpensive, simple, specialised, and lethal platforms – and to risk them deliberately in battle. If constructed to carry a crew, many of these desirable attributes fall by the wayside. What, after all, is the difference between today’s multipurpose warship and a crewed autonomous vessel?

The medium and large unmanned surface and underwater vessels the USN needs for a high-end fight in the western Pacific are vastly different from the small, inexpensive, aerial drones that hunt on the Ukraine steppe, or even on the Black Sea. Given the harsh operating environment, distances, and intended targets, autonomous surface and subsurface platforms will have to be relatively large and expensive. Nevertheless, autonomy creates an opportunity to build lethal systems for a fraction of the cost of a crewed vessel. The new Constellation-class frigate (if built) will cost approximately $1.2 bn dollars each.[21] It should be possible to build 30 or more large, armed autonomous vessels for the same price. Thirty such USVs, each armed with 10 antiship cruise missiles, would significantly magnify the combat capability of a traditional three-ship surface action group.

Autonomy can decrease costs, increase numbers, and enhance lethality while providing commanders with assets that can be expended in battle. Integrated into the Navy’s existing force structure and employed with dexterity, it might offer a better way to deter an opponent who possesses a superior number of multipurpose warships. The essence of autonomous vessels will never manifest, however, if they are built to carry crews.

This article originally appeared in the U.S. Naval Institute Proceedings. Copyright U.S. Naval Institute. Reprinted with permission. For more great content from the U.S. Naval Institute, visit www.usni.org.

References

[1] Geoffrey Wawro, The Franco-Prussian War: The German Conquest of France in 1870–1871 (Cambridge, UK: Cambridge University Press, 2003).

[2] Colin S. Gray, War, Peace and International Relations: An Introduction to Strategic History (London: Routledge, 2012), 19. Gray also noted, “new capabilities inexorably bring new vulnerabilities,” Colin Gray, “The Changing Nature of Warfare?” Naval War College Review XLIX, no. 2 (Spring 1996): 14.

[3] https://www.usni.org/press/books/innovating-victory ; Vincent P. O’Hara and Leonard R. Heinz, Innovating Victory: Naval Technology in Three Wars (Annapolis, MD: Naval Institute Press, 2022).

[4] David Kirichenko, “Drone superpower: Ukrainian Wartime Innovation Offers Lessons for NATO,” Atlantic Council, 13 May 2025.

[5] https://spectrum.ieee.org/sea-drone Bryan Clark, “Sea Drones in the Russia-Ukraine War Inspire New Tactics,” IEEE Journal, 10 July 2024.

[6] Masaao Dahlgren and Lachlan MacKenzie, “Ukraine’s Drone Swarms Are Destroying Russian Nuclear Bombers. What Happens Now?” Center for Strategic & International Studies, 4 June 2025.

[7] Maziar Motamedi, “How Israel Launched Attacks from Inside Iran to Sow Chaos During War,” Al Jazerra, 26 June 2025.

[8] David M. Kaplan, ed., Readings in the Philosophy of Technology (Lanham, MD: Rowman & Littlefield, 2009).

[9] Rauno Huttunen and Leena Kakkon, “Heideggger’s Critique of the Technology and the Education Ecological Imperative,” Educational Philosophy and Theory 54, no. 5 (May 2022): 630–42.

[10] Martin Heidegger, The Question Concerning Technology and Other Essays (New York: Harper Torchbooks, 1977), 14.

[11] Robert V. Kozinets, “Technology/Ideology: How Ideological Fields Influence Consumers’ Technology Narratives,” Journal of Consumer Research 34, no. 6 (April 2008): 865–81.

[12] Jeffrey E. Kline, James A. Russell, and James J. Wirtz, “The Navy’s Generational Challenge,” Survival (August/September 2022), 123–36.

[13] On the physics involved see CAPT Talbot Manvel, USN (Ret.), “Aircraft Carriers: Bigger is Better,” U.S. Naval Institute Proceedings 146, no. 9 (September 2020). https://www.usni.org/magazines/proceedings/2020/september/aircraft-carriers-bigger-better

[14] James J. Wirtz, Jeffrey E. Kline and James A. Russell, “A Maritime Conversation with America,” Orbis (Spring 2022), 179.

[15] Hope Hodge Seck, “The Navy to Simplify Drone Ship Plans, Focus on Containerized Payloads that Look Alike,” The War Zone, 17 January  2025.

[16] Craig L. Symonds, “The Navy Saved Our Hides,” Naval History 28, no. 3 (June 2014). https://www.usni.org/magazines/naval-history-magazine/2014/june/navy-saved-our-hides

[17] Samuel Eliot Morison, The Invasion of France and Germany: History of United States Naval Operations in World War II, vol XI (Boston: Little Brown, 1947), 142–49.

[18] Vincent P. O’Hara and Leonard R. Heinz, Innovating Victory: Naval Technology in Three Wars (Annapolis, MD: Naval Institute Press, 2022), 52–80.

[19] Seck, “The Navy to Simplify Drone Ship Plans.”

[20] Congressional Budget Office, “The Cost of the Navy’s New Frigate,” October 2020.

[21] https://news.usni.org/2020/10/14/cbo-report-on-cost-of-the-navys-ffgx-program