Debate on Autonomous Weapons: A Misguided Starting Point

The discussion on autonomous weapons often begins on shaky ground.

Critics question whether the U.S. should let machines take lives, while proponents concern themselves with the potential risks of falling behind adversaries who may utilize these technologies. Both camps see autonomous lethality as a moral dilemma that necessitates an entirely new system of governance.

The reality is that the U.S. has already laid the groundwork for deploying autonomous systems, independent of whether the public discourse achieves a clear resolution.

In fact, the military has utilized a structured approach for decades, which can be intrinsically linked to the use of autonomous systems. Understanding this framework could significantly enhance the public debate.

Force management is handled via a systematic structure categorized by the status of weapons, familiar to air defense operators and ground commanders alike. The three classifications include: “Keep Arms,” which authorizes engagement solely in self-defense or under specific commands; “Tight Weapons,” permitting engagement only against clearly hostile targets; and “No Weapons Use,” which allows engagement against targets whose status isn’t entirely clear.

Commanders determine these statuses based on the mission, perceived threats, and surrounding circumstances. Different units can operate under varying statuses depending on the scenario. This system already adapts lethal power to fit specific situations. The commander sets the control decisions, so there’s no need to seek individual soldier approval for every action.

This framework can be readily applied to the issue of autonomous weapons.

The argument that machines lack the contextual judgment to distinguish between threats, civilians, and bystanders only holds in environments where making such distinctions is especially challenging—precisely the scenarios that existing arms control frameworks are designed to address.

The contexts of the Taiwan Strait and urban areas like downtown Tehran are not interchangeable, and they certainly shouldn’t be governed by the same overarching framework.

Take, for instance, the operations in the Taiwan Strait. Autonomous systems there are focused on naval engagements within conflict zones, where civilian presence is minimal. All vessels fitting a specific profile are considered hostile, so the distinction between threats remains straightforward. The environment is structured and the military objectives are clear. Failure to act can lead to the loss of American vessels and personnel from enemy strikes that outpace human pilot reactions.

Thus, deploying an armed or even heavily armed autonomous posture is justifiable in such scenarios, reflecting the logic that supports human air defense strategies.

Conversely, the same autonomous systems working in densely populated urban areas, such as downtown Tehran, would require weapons to be restrained and necessitate human authorization after each engagement. The surroundings dictate how these systems should behave, and a already-established framework can guide those actions.

The Department of Defense is actually beginning to incorporate these existing rules into current policies. Directive 3000.09, updated in January 2023, mandates that autonomous and semi-autonomous weapon systems allow commanders to exercise appropriate human judgment regarding the use of force. It also requires that each engagement aligns with the intent of the commander and remains limited by time and geographic boundaries.

This directive acknowledges that the level of human control should vary based on the specific system and mission at hand.

What the directive has yet to tackle, and what public discussion hasn’t clarified, is how to align these variations with the established arms control language familiar to military personnel, so that commanders, policymakers, and the general public can grasp the issue through a lens that has been in use for years.

Embracing this perspective calls for a trust in the military’s capability, crucial for public authorities. The notion that the U.S. cannot trust commanders with autonomous lethal force often borders on exaggeration.

We already rely on those same commanders to manage human lethal forces within the same framework. Under this system, civilian casualties occur when commanders make incorrect decisions.

Autonomous systems governed by a similar rationale bear the same level of accountability. A commander who opts for a restrained posture with an autonomous unit will face the same consequences as one who does so with a team of human-operated interceptors.

It’s essential for authorities overseeing autonomous forces to clearly establish this policy, rather than allowing it to evolve on a case-by-case basis through procurement choices and after-action reviews.

In general, militaries should publicly affirm that autonomous systems work under the established arms control classifications set by responsible commanders. The specific status assigned to any system should align with the challenges of identifying its operational environment. The most permissive stance should only be available in scenarios where distinguishing between threats and non-threats is relatively straightforward.

Such clarity offers two key benefits absent from the current ambiguous debate. First, it provides commanders with a well-defined, recognizable terminology regarding the governing system—terminology they would otherwise have to navigate without doctrinal guidance. Second, an identifiable arms control status offers transparency and accountability rather than relying on an opaque algorithm that lacks a clear owner.

Ultimately, the alternative isn’t a future devoid of autonomous weapons. Adversaries are already developing these systems, and technology is spreading. The U.S. will roll out its own autonomous systems, independent of whether public discourse reaches a conclusion.

A lack of an explicit framework—to instead function under vague guidelines—forces operators to make critical decisions on the fly rather than allowing commanders to manage objectives within a well-established system in place for decades.

Military entities are experienced in applying lethal force, and this framework is credible, familiar, and holds leaders accountable. The challenge lies in applying it purposefully to new autonomous systems rather than assuming such technologies need entirely new ethical guidelines.