Lethal autonomous weapon

Definitions

In weapons development, the term "autonomous" is somewhat ambiguous and can vary substantially between different scholars, nations and organizations. There is no definition of lethal autonomous weapon systems that is generally agreed upon among different countries.

Some of the definitional problems are noted in the UN Secretary-General's report to the General assembly. For instance, some observers think that the word "lethal" should not be used in discussing these advanced weapons because lethality is "an effect of the manner in which a weapon is used" rather than a necessarily innate quality of the weapon." The meaning of "autonomous" also is disputed inasmuch as there are varying degrees of autonomy within existing systems that are commonly thought of as autonomous.

The official United States Department of Defense Policy on Autonomy in Weapon Systems (Department of Defense Directive 3000.09) defines an Autonomous Weapon (note that the US tends to refer to "weapon system," not "weapons system") System as one that "...once activated, can select and engage targets without further intervention by a human operator." Heather Roff, a writer for Case Western Reserve University School of Law, describes autonomous weapon systems as "... capable of learning and adapting their 'functioning in response to changing circumstances in the environment in which [they are] deployed,' as well as capable of making firing decisions on their own."

The United Nations Group of Governmental Experts on Emerging Technologies in the Area of Letha Autonomous Weapons Systems (GGE) has defined a LAWS as "an integrated combination of one or more weapons and technological components that enable the system to identify and/or select, and engage a target, without intervention by a human user . . .

The United States Department of Defense definition of an AWS is more extensive than that of the GGE. DoD Directive 3000.09 is similar to that of the GGE in that it defines an AWS as "a weapon system that, once activated, can select and engage targets without further intervention by an operator." However, the directive goes on to stipulate that its definition "includes, but is not limited to, operator-supervised autonomous weapon systems that are designed to allow operators to override operation of the weapon system, but can select and engage targets without further operator input after activation." The British Ministry of Defence states "Whilst definitions can vary, the key difference is that an automated system is capable of carrying out complicated tasks but is incapable of complex decision-making, whereas an autonomous system is capable of deciding a course of action without depending on human oversight and control." Some nations (such as France) take a restrictive view of what autonomy consists of, defining an AWS "as a fully autonomous' weapon system, meaning that it is 'capable of changing its own rules of operation, particularly as regards target engagement beyond a determined framework of use'" (emphasis added)

On the other hand, scholars such as Peter Asaro and Mark Gubrud believe that any weapon system that is capable of releasing a lethal force without the operation, decision, or confirmation of a human supervisor can be deemed autonomous. Nevertheless, a wide array of nations and organizations, such as the US Department of Defense, the International Committee of the Red Cross, Human Rights Watch, and AWS critics agree that the essential quality of autonomous weapon systems is that they, "once activated, are capable of selecting and engaging targets without contemporaneous human input.

The meaning of autonomy in the weapons context has been made less precise in recent years by some commentators using it interchangeably with Artificial Intelligence (AI). One observer has noted that "reliance on remotely piloted capabilities, uncrewed vehicles, and loitering munitions, with ever-increasing degrees of autonomy (AI enabled or not) . . . are further blurring the lines between AI, autonomy, and military systems . . ."

3 more chapters are free with a Lably account.

Automatic defensive systems

Some definitions of autonomous weapon systems are broad enough to include land mines and naval mines, simple automatically-triggered lethal weapons that have been in use for centuries.

Some current examples of LAWs are automated "hardkill" active protection systems, such as a radar-guided close-in weapon systems (CIWS) used to defend ships that have been in use since the 1970s (e.g., the US Phalanx CIWS). Such systems can autonomously identify and attack oncoming missiles, rockets, artillery fire, aircraft, and surface vessels according to criteria set by the human operator. Similar systems exist for tanks, such as the Russian Arena, the Israeli Trophy, and the German AMAP-ADS. Several types of stationary sentry guns, which can fire at humans and vehicles, are used in South Korea and Israel. Many missile defence systems, such as Iron Dome, also have autonomous targeting capabilities.

The main reason for not having a "human in the loop" in these systems is the need for rapid response. They have generally been used to protect personnel and installations against incoming projectiles.

Autonomous offensive systems

According to The Economist in 2018, as technology advances, applications of uncrewed undersea vehicles could include mine clearance, mine-laying, anti-submarine sensor networking in contested waters, patrolling with active sonar, resupplying manned submarines, and becoming low-cost missile platforms.

In 2017 the Russian Federation was developing artificially intelligent missiles, drones, unmanned vehicles, military robots and medic robots. In 2018, the U.S. Nuclear Posture Review alleged that Russia was developing a "new intercontinental, nuclear-armed, nuclear-powered, undersea autonomous torpedo" named "Status 6".

Israeli Minister Ayoob Kara stated in 2017 that Israel is developing military robots, including ones as small as flies.

In October 2018, Zeng Yi, a senior executive at the Chinese defense firm Norinco, gave a speech in which he said that "In future battlegrounds, there will be no people fighting", and that the use of lethal autonomous weapons in warfare is "inevitable". In 2019, US Defense Secretary Mark Esper lashed out at China for selling drones capable of taking life with no human oversight.

As of 2020, DARPA was working on making swarms of 250 autonomous lethal drones available to the American military. The US Navy is developing unmanned surface vehicles, also called sea drones, including Ghost Fleet Overlord, with plans to equip them with weapons and with the potential to use them semi-autonomously.

In 2020 a Kargu 2 drone hunted down and attacked a human target in Libya, according to a report from the UN Security Council's Panel of Experts on Libya, published in March 2021. This may have been the first time an autonomous killer robot armed with lethal weaponry attacked human beings.

In May 2021 Israel conducted an AI-guided combat drone swarm attack in Gaza.

In the Russo-Ukrainian war, Ukraine has developed advanced drones with integrated artificial intelligence for a range of drone warfare purposes, including to attack infrastructure in Russia, although as of May 2026, Al Jazeera reported that humans remain in control of operation.

In June 2026, a senior figure in the Ukraine defense industry told New Scientist that, two years earlier, "fully autonomous drones with no human oversight" had killed soldiers on the battlefield for the first time. The attack involved 10 AI-controlled "Terminator" drones, which were said to have killed "a couple" of Russian soldiers, and to have additionally targeted a Russian tank. A Ukrainian drone-maker described the weapon as follows: "We just launch it and we know everything will be dead – everything that will be found there in this particular area will be dead... There is no connection to the drone at all, you cannot see the video, nothing... Everything it sees will be killed."

Ethical and legal issues

Degree of human control

The increasing use of AWS raises numerous ethical and legal issues. David Kirichenko has asserted: “This is the Oppenheimer moment of our generation.” He warns that autonomous drones could make mass killing nearly effortless and views an AWS arms race as distinctly possible. Backing-up this claim, Kirchenko notes that drones now are responsible for about seventy percent of all casualties in the Russo-Ukrainian war. Three classifications of the degree of human control of autonomous weapon systems were laid out by Bonnie Docherty in a 2012 Human Rights Watch report: 1)human-in-the-loop: a human must instigate the action of the weapon (i.e. the weapon is not fully autonomous); 2)human-on-the-loop (a human may abort an action; 3)human-out-of-the-loop (no human action is involved. More recently, the phrase "meaningful human control" (along with “human control," "appropriate human judgment," and "human involvement") has been used increasingly in discussions concerning limits on the use of automated weapon systems. Among the retained elements of human control that have been suggested as necessary are: 1)“sufficient information, including on the capabilities of the weapons system and the operational context, to ensure compliance with international law; 2) the ability to exercise their judgment to the extent required by international humanitarian law; 3) the ability to limit the types of tasks and targets."

Standard used in US policy

Department of Defense Directive 3000.09 states that "Autonomous … weapons systems shall be designed to allow commanders and operators to exercise appropriate levels of human judgment over the use of force."

However, as noted in the Bulletin of the Atomic Scientists, the policy requires that autonomous weapon systems that kill people or use kinetic force, selecting and engaging targets without further human intervention, be certified as compliant with "appropriate levels" and other standards, not that such weapon systems cannot meet these standards and are therefore forbidden. "Semi-autonomous" hunter-killers that autonomously identify and attack targets do not even require certification.

Deputy Defense Secretary Robert O. Work said in 2016 that the Defense Department would "not delegate lethal authority to a machine to make a decision", but might need to reconsider this since "authoritarian regimes" may do so. In October 2016 President Barack Obama stated that early in his career he was wary of a future in which a US president making use of drone warfare could "carry on perpetual wars all over the world, and a lot of them covert, without any accountability or democratic debate". In the US, security-related AI has fallen under the purview of the National Security Commission on Artificial Intelligence since 2018.

On October 31, 2019, the United States Department of Defense's Defense Innovation Board published the draft of a report outlining five principles for weaponized AI and making 12 recommendations for the ethical use of artificial intelligence by the Department of Defense that would ensure a human operator would always be able to look into the black box and understand the kill-chain process. A major concern is how the report will be implemented.

Possible violations of ethics and international law

Stuart Russell, professor of computer science from University of California, Berkeley stated the concern he has with LAWs is that his view is that it is unethical and inhumane. The main issue with this system is it is hard to distinguish between combatants and non-combatants. However, improvements in AWS, especially with the use of AI, can ameliorate some of these concerns. Nathan Wood has noted that "there are a host of cases where specific autonomous systems can reliably distinguish between military targets and protected persons and objects, and in some cases already are doing so.

There is concern by some economists and legal scholars about whether LAWs would violate International Humanitarian Law, especially the principle of distinction, which requires the ability to discriminate combatants from non-combatants, and the principle of proportionality, which requires that damage to civilians be proportional to the military aim. This concern is often invoked as a reason to ban "killer robots" altogether - but it is doubtful that this concern can be an argument against LAWs that do not violate International Humanitarian Law. Wood has asserted that while it would be wrong to use AWS in some circumstances, in others AWS "are likely to be superior, both morally and legally, to conventional alternatives."

A 2021 report by the American Congressional Research Service states that "there are no domestic or international legal prohibitions on the development of use of LAWs," although it acknowledges ongoing talks at the UN Convention on Certain Conventional Weapons (CCW).

LAWs are said by some to blur the boundaries of who is responsible for a particular killing. Philosopher Robert Sparrow argues that autonomous weapons are causally but not morally responsible, similar to child soldiers. In each case, he argues there is a risk of atrocities occurring without an appropriate subject to hold responsible, which violates jus in bello. Thomas Simpson and Vincent Müller argue that they may make it easier to record who gave which command.

Campaigns to ban LAWs

The possibility of LAWs has generated significant debate, especially about the risk of "killer robots" roaming the earth in the near or far future. The group Campaign to Stop Killer Robots formed in 2013. In July 2015, over 1,000 experts in artificial intelligence signed a letter warning of the threat of an artificial intelligence arms race and calling for a ban on autonomous weapons. The letter was presented in Buenos Aires at the 24th International Joint Conference on Artificial Intelligence (IJCAI-15) and was co-signed by Stephen Hawking, Elon Musk, Steve Wozniak, Noam Chomsky, Skype co-founder Jaan Tallinn and Google DeepMind co-founder Demis Hassabis, among others.

According to PAX For Peace, one of the founding organisations of the Campaign to Stop Killer Robots, fully automated weapons (FAWs) will lower the threshold of going to war as soldiers are removed from the battlefield and the public is distanced from experiencing war, giving politicians and other decision-makers more space in deciding when and how to go to war. They warn that once deployed, FAWs will make democratic control of war more difficult, something that author of Kill Decision (a novel on the topic) and IT specialist Daniel Suarez also warned about: according to him it might recentralize power into very few hands by requiring very few people to go to war.

Persons protesting the development of LAWS have created websites aimed at publicizing what they view as the undesirable effects of research into the application of artificial intelligence to weapons design. On these websites, news about ethical and legal issues are constantly updated for visitors to recap with recent news about international meetings and research articles concerning LAWs. In December 2022, a vote of the San Francisco Board of Supervisors to authorize San Francisco Police Department use of LAWs drew national attention and protests. The Board reversed this vote in a subsequent meeting.

The Holy See has called for the international community to ban the use of LAWs on several occasions. In November 2018, Archbishop Ivan Jurkovic, the permanent observer of the Holy See to the United Nations, stated that "In order to prevent an arms race and the increase of inequalities and instability, it is an imperative duty to act promptly: now is the time to prevent LAWs from becoming the reality of tomorrow’s warfare." The Church has expressed concern that these weapons systems have the capability to irreversibly alter the nature of warfare, create detachment from human agency and put in question the humanity of societies. However, the Holy See’s call for a ban on lethal autonomous weapons refers only to those weapons whose critical functions, such as target selection and engagement, are under the machine’s control instead of a human being’s. In the Holy See’s view “a machine can be thus considered as under ‘significant human control’ or MHC [Meaningful Human Control] if: (1) there is a degree human supervision at least capable of calling off a particular sortie; (2) all courses of action are ‘well known’ to the operator; (3) the environment is 'perfectly circumscribed and known." Apostolic Nuncio Ettore Balestrero reiterated the Holy See's position on LAWS in 2024.

As of 29 March 2019, the majority of governments represented at a UN meeting to discuss the matter favoured a ban on LAWs. A minority of governments, including those of Australia, Israel, Russia, the UK, and the US, opposed a ban. The United States has stated that autonomous weapons have helped prevent the killing of civilians.

Regulation without banning

Another approach focuses on regulating the use of autonomous weapon systems in lieu of a ban. Military AI arms control will likely require the institutionalization of new international norms embodied in effective technical specifications combined with active monitoring and informal ('Track II') diplomacy by communities of experts, together with a legal and political verification process. In 2021, the United States Department of Defense requested a dialogue with the Chinese People's Liberation Army on AI-enabled autonomous weapons but was refused.

Under the framework of the Convention on Certain Conventional Weapons, states have discussed lethal autonomous weapon systems since 2014. In 2016, the treaty's states parties established an open-ended Group of Governmental Experts on Lethal Autonomous Weapons Systems to continue those discussions. The discussions have addressed international humanitarian law, accountability, possible prohibitions and regulations, and the extent of human control required over AI-enabled weapons. As of 2026, the GGE had reached a provisional consensus on twenty-two points "for the purpose of advancing its work on a set of elements of an instrument, without prejudging its nature, and other possible measures, in accordance with the Group’s mandate." It has been suggested that there are three approaches to creating a legally binding instrument governing LAWS: 1) the prohibition of particular types of autonomous weapons (such as those targeting humans); 2) imposing obligations regarding AW use (such as requiring meaningful human control); and 3) mandating due diligence requirements in the use of autonomous weapons.

A summit of 60 countries was held in 2023 on the responsible use of AI in the military. On 22 December 2023, a United Nations General Assembly resolution was adopted to support international discussion regarding concerns about LAWS. The vote was 152 in favor, four against, and 11 abstentions.