Research on Legal Issues Related to the Utilization of Unmanned Combat Platforms

【Lecture Hall】

Unmanned combat platform refers to a type of weapon platform that is unmanned, fully operated by remote control or operated autonomously according to pre-programmed procedures, and carries offensive or defensive weapons to carry out combat missions. It mainly includes unmanned combat aircraft and unmanned anti-submarine Warfare or anti-mine warfare submersibles and unmanned combat vehicles.

Under the existing technical conditions, unmanned combat platforms generally cannot automatically select and attack targets, and are still in the "remote control" or "semi-autonomous" development stage.

However, with the development of artificial intelligence technology and the improvement of the ability to independently select and strike targets, unmanned combat platforms have a tendency to accelerate the development of autonomous weapons.

Autonomous weapons refer to the ability to search and identify without human intervention and control, and then use lethal force to attack targets including humans. Generally, they have "autonomy of attack" and "free from human interference or manipulation". Elements.

It can be seen that in the current practice of various countries, not all unmanned combat platforms are autonomous weapons.

The legal consensus of the international community on unmanned combat platforms

  At present, the international community does not have a unified definition of unmanned combat platforms, nor has it formed a convention or regulated it in other legislative forms.

However, since 2014, the United Nations has held a dedicated discussion on lethal autonomous weapon systems at the "Convention on Certain Conventional Weapons" meeting held in Geneva, Switzerland every year.

In 2017, the meeting established the Government Expert Group (GGE) to discuss important issues such as the definition, characteristics, human-computer interaction, and application of international humanitarian law of autonomous weapons, and basically reached the following consensus at the legal level: First, the existing international humanitarian law It still applies to the control of autonomous weapons. Autonomous weapons must strictly abide by the basic provisions of international humanitarian law, including the principle of distinction, proportionality, military necessity, and human rights; the second is to deploy unmanned combat weapon systems in armed conflicts, and their responsibilities It depends entirely on the state actors that use weapon systems; third, a completely autonomous weapon platform that has no human intervention and has the power to determine life and death is absolutely unacceptable. In other words, since autonomous weapons are manufactured by humans, they must be subject to human control.

As the early form of autonomous weapons, unmanned combat platforms should also comply with the aforementioned humanitarian requirements.

Challenges of unmanned combat platforms to international humanitarian law

  As unmanned weapon systems such as unmanned submarines, unmanned aerial vehicles, battlefield robots continue to increase their anthropomorphism, unmanned combat platforms pose challenges to the law of armed conflict and international humanitarian law when performing attack tasks, which are mainly reflected in the following aspects :

  Rules for the use of impact force.

The United Nations Charter clearly prohibits the use of force and regulates the special circumstances of the use of force.

However, with the widespread use of unmanned combat platforms, combatants can fight in indirect combat away from the battlefield, which makes the conditions and standards for the use of force blurred, and reduces the political risk and moral pressure of the use of force. The manipulator’s mentality of staying out of the incident This may lead to randomness in the use of force.

After the "September 11" incident, the United States often sent drones at will to search and strike countries where terrorists might be hiding, rudely trampling on the sovereignty of other countries.

In addition, due to the "swarming" use of drones and unmanned boats, it is very suitable for pre-emptive attacks on enemy nuclear and non-nuclear mobile missile launchers, ballistic missile nuclear submarines and their auxiliary facilities (such as reconnaissance, surveillance and early warning systems, etc.) , The adversary may take the worst countermeasures when it is unable to determine the attack path and target, that is, implement the “attack and launch” strategy of nuclear retaliation. This will greatly reduce the threshold for the use of nuclear weapons and cause national security to be systematically affected. The possibility of overall threats is rising, and the difficulty of establishing strategic mutual trust is getting higher and higher. Some countries, especially major powers, have a decline in their recognition of the international system and rules of international law.

  The principle of shock distinction and the principle of proportionality.

Although the unmanned combat platform has a certain degree of accuracy in judging humans and non-humans, the robots will only strictly execute human pre-set procedures and brutally kill all locked targets. It treats civilians and combatants as well as among civilians. It is difficult to distinguish the combatants of, and it is impossible to accurately distinguish the wounded from the prisoners, which can easily lead to manslaughter and accidental injury.

At the same time, unmanned combat platforms can attack without interruption without considering too many factors. It is easy to cause indiscriminate unnecessary casualties while achieving results. This will undoubtedly increase the real danger of innocent civilians and civilian objects on the battlefield.

For example, in February 2010, during a mission in Afghanistan, a "carnivorous animal" unmanned attack aircraft "provided inaccurate and unprofessional reports" and mistakenly attacked three passenger cars full of civilians, resulting in 23 The death of an Afghan civilian is clearly a serious violation of the basic principles of international humanitarian law.

According to incomplete statistics from the New America Foundation, the number of U.S. drone attacks in Pakistan from 2004 to 2013 was 357, and about 20% of civilians and unidentified persons were innocent victims.

  Cause the responsibility mechanism of international law to face difficulties.

Responsibility and accountability are the core content of international humanitarian law, involving the national responsibility for illegal research and development, acquisition, and application of unmanned combat platforms, and the international criminal responsibility of relevant individuals for violating the laws of war.

If the unmanned combat platform launches a wrong attack, injures (kills) civilians by mistake, or violates the principles of international humanitarian law, how to impose liability becomes a difficult question.

Who will be responsible for the decisions made by the platform?

Who is responsible for the violations of international humanitarian law implemented by the platform?

Is it the programmer, the manufacturer, or the commander who deployed the weapon system?

The core key technology of unmanned combat platforms is artificial intelligence, and artificial intelligence has two important shortcomings: inexplicability and the inevitability of algorithm loopholes.

Unexplainability is mainly derived from its own technical logic: artificial intelligence based on big data learning depends on probability theory rather than causality in methodology.

Based on the inherent deficiencies of artificial intelligence, most of the potential risks of unmanned combat platforms are difficult to attribute to individual subjective intentions, which makes it difficult to pursue international criminal responsibility for violations of international humanitarian law in the application of unmanned combat platforms.

  The implementation of legal reviews has become increasingly complex.

As a new type of warfare means and method, the use of unmanned combat platforms should be reviewed by international humanitarian law.

Based on the unpredictability of the performance and effect of unmanned combat platforms, the legal review of unmanned combat platforms under the background of artificial intelligence has become more complicated and difficult.

In February 2020, the US Department of Defense officially launched the "Artificial Intelligence Ethical Principles", proposing that military applications of artificial intelligence should follow the five principles of "responsibility, fairness, traceability, reliability, and control".

The world's first military artificial intelligence ethics principle introduced by the US Defense Innovation Council is somewhat different from the principles advocated by the United Nations, and it is also different from the EU and China's basic stand against "artificial intelligence militarization".

What deserves our vigilance is that the United States is trying to evolve its own ethical standards into the fundamental principles of international rules in this field, to obtain the final interpretation of the militarization of artificial intelligence, and to lead its loyal allies that have been nurtured by its guiding principles and safety procedures for a long time. Formulate a series of international laws that are conducive to the interests of the United States.

Legal issues related to underwater unmanned combat platforms

  Underwater unmanned combat platform is a type of underwater unmanned submarine vehicle (UUV for short). It refers to unmanned operation, through preset procedures or remote control, used for attack and other combat and lethal underwater navigation power Device.

So far, the status of military underwater unmanned vehicles in international law is unclear.

  水下无人潜航器具有多元法律属性。水下无人潜航器种类繁多,因操控方式不同其法律属性引发国际法争议:水下无人潜航器是属于《联合国海洋法公约》(简称《公约》)语境下的军舰、潜艇,或是海商法语境下的船舶属具、船舶,抑或是国际人道法语境下的自主武器系统?综合大多数学者的观点,笔者认为,不同类别的水下无人潜航器应当具有不同的法律属性。首先,依据《公约》对“军舰”定义,无人潜航器既未明确隶属武装部队,也未被列入现役军舰名册,更不可能在潜航器上配备船员,因而不能定性为“军舰”。其次,按照操作方式的不同,从海洋法和海商法角度,船基遥控型无人潜航器不具备独立的船舶法律地位,更适于定性为母船的附属设施即“船舶属具”;而自主型、程控型和岸基遥控型无人潜航器则可认定为“执行军事活动的船舶”,受相关海事法律的调整。2017版《美国海军海上行动法指挥官手册》将无人潜航器视为“其他类型的海军舰船”,并不完全准确。再次,武器型无人潜航器即水下无人作战平台属于自主武器系统,不具备船舶法律地位。近几年来,武器型无人潜航器的技术发展迅速,其自主决策攻击能力大大提升,比如2018年7月完成测试的俄罗斯海军“波塞冬”核动力无人潜航器(以及俄罗斯确认正在研制的“海神”核动力无人潜航器)技术先进,最具有代表性。据报道,“波塞冬”无人潜航器可以从两种不同级别的核潜艇上发射,既可发射常规鱼雷,也可装备核弹头进行核打击,它能在锁定目标后进行自动洲际潜航,发射的鱼雷下潜深度和航行速度均超过同类武器。武器型无人潜航器的技术发展及其自主选择和打击目标能力的提升,使其具备自主武器的基本要素。虽然这种无人潜航器本身并不违反现行国际人道法,但是作为一种自主武器系统,攻击型无人潜航平台与其他自主武器一样,可能带来的国际人道法问题和战争伦理问题不容忽视,因而应纳入《特定常规武器公约》会谈机制的探讨范畴。

  Does the underwater unmanned submarine enjoy immunity?

At present, the U.S. Navy has extensively used unmanned submarine vehicles to detect military intelligence such as marine geographic information of other countries.

U.S. officials believe that unmanned underwater vehicles enjoy sovereign immunity, and believe that “no matter whether they are in international waters or sovereign waters, these ships cannot be arrested and searched by other countries”, that is, they are free from arrest, seizure, interference, etc. from coastal countries. "Absolute immunity" of measures.

The author believes that unmanned submarines that do not have the attributes of warships and government ships do not enjoy immunity.

At present, the status of unmanned submersibles defined as ships in international law has become a general trend.

The "Convention" stipulates that only ships and warships owned or operated by one country and dedicated to government non-commercial services can enjoy immunity when operating at sea.

It has been analyzed that the unmanned submarine does not have the external mark to identify the nationality of the warship, and it is not equipped with crew members who obey the discipline of the regular armed forces, so it cannot be classified as a "warship."

Government ships refer to ships "used for non-commercial purposes" and "services to the government that can be clearly identified".

Underwater unmanned submersibles engaged in "military oceanographic survey missions" do not serve the national government but the navy. Therefore, unmanned submarines do not meet the definition of government ships.

  Whether unmanned submarine vehicles used for military purposes enjoy immunity is not yet clearly regulated by international treaties. The so-called view of sovereign immunity by the United States is not invulnerable.

What is certain is that unmanned submarine vehicles of other countries entered our exclusive economic zone without consent to conduct military surveys and other activities that endanger China's security, which violated the "Convention" and relevant Chinese laws.

For the unidentified objects found in the seas under my jurisdiction, for safety reasons, I have the right to organize the first salvage, the relevant departments identify and verify, properly handle and respond, and play a proper control and warning role.

In addition, if it is a weapon-type unmanned submarine, it does not have the legal status of a ship and does not enjoy immunity.

  Immunity is not a right that can be done arbitrarily.

The basic principles of modern international law require countries to respect each other's sovereignty and security.

Under general international law, immunity mainly refers to immunity from jurisdiction and property enforcement. It does not exempt foreign warships and government ships from illegal responsibilities. Of course, it does not exempt countries from taking certain measures in accordance with the law, such as expulsion, surveillance, verification and identification.

Even if a country’s warships and government ships enjoy immunity, it does not mean that they can act arbitrarily regardless of international law and the domestic laws of coastal countries, nor can they use them to engage in illegal acts that endanger the sovereignty and security of other countries, because this violates the purpose of immunity. And purpose.

The "Convention" on the immunity of warships and government ships is stipulated in Article 32, Article 95-96 and Article 236.

Article 32 stipulates that this immunity does not include the circumstances stipulated in subsection A of the territorial sea regime and Articles 30-31; the "complete immunity" stipulated in Articles 95-96 is limited to the high seas; Article 236 stipulates sovereign immunity Applicable to the "Provisions of this Convention on the Protection and Preservation of the Marine Environment".

Article 58 paragraphs 1 and 2 of the "Convention" introduce part of the high seas regime into the exclusive economic zone, such as freedom of navigation and overflight, but the premise is "under the restrictions of the relevant provisions of this Convention" "as long as it does not conflict with this part "Should give due consideration to the rights and obligations of coastal states, and shall abide by the laws and regulations that are not inconsistent with this part made by coastal states in accordance with the provisions of this Convention and other rules of international law."

It can be seen that immunity is not a substitute for infringement. There is no legal basis for enlarging the content of immunity to "doing whatever you want without control."

Thoughts and suggestions

  Under the current international law framework, the unmanned combat platform itself does not violate the relevant provisions of international law.

At present, the development of unmanned combat platforms has become an unstoppable trend. We must not only actively promote its rational use in the military field, but also discuss the use of military unmanned systems.

In response to the ethical principles of military artificial intelligence proposed by the United States, China proposes on the basis of seeking common ground while reserving differences, and continues to advocate and promote the adoption of the protocol on autonomous weapons by the United Nations to enhance China's international voice in the field of military unmanned technology.

  Whether the use of unmanned combat platforms complies with international humanitarian law depends on many factors such as the tasks performed by the platform, applicable links, and intent to attack targets.

In practice, the technical risks can be avoided through proper deployment and preventive measures.

Take an attack drone as an example. It is controlled by combatants and issued attack orders. According to the principle of distinction, it can find out whether the target can be attacked. According to the principle of proportionality, it can determine whether to launch an attack. It takes civilian casualties and damage to civilian objects. Minimization and other measures to ensure that unmanned combat platform applications are targeted at specific military targets and will not cause excessive harm or unnecessary suffering.

  Promote the establishment of a legal review mechanism for new weapons.

Unmanned combat platforms are a collection of weapons, combat means, and methods. A legal review mechanism should be established in accordance with Article 36 of the First Additional Protocol to the Geneva Conventions, and the review criteria and scope of review should be clarified.

Establish "meaningful human control principles" in the technological and legal development of unmanned combat platforms, emphasize remote control by commanders, formulate various disposal plans under complex battlefield conditions from the technical level, and provide emergency backup and security measures , To ensure that the final decision-making power is in the hands of people, strengthen the control of lethal autonomous weapon systems, and reasonably avoid international disputes.

  Promote the international community to establish and improve accountability mechanisms.

Improve the national responsibility system for violations of international humanitarian law due to the development and use of unmanned combat platforms, and improve the international criminal responsibility system for combatants, their commanders, and R&D personnel for violations of international humanitarian law by unmanned combat platforms in armed conflicts.

Generally speaking, the commander is the primary option for unmanned combat platforms to generate personal criminal liability.

The relevant regulations of the US Department of Defense can be used as a reference: personnel who directly apply, authorize or operate autonomous weapon systems must take appropriate care and use them in accordance with war laws and treaties, as well as weapon system safety rules and operational rules.

This means that if the commander knows or should know that the autonomous weapon system put into service will violate international law, he will bear personal international criminal responsibility for the act.

(Author: Mao Guohui, Professor of National University of Defense Technology)

Keywords: unmanned, unmanned aerial vehicles, convention on certain conventional weapons, human-computer interaction, remote control, autonomy of attack, that is, the u.s., of course, and about, lecture hall, military necessity, generally, they have, interference