Rules of Engagement as a Security Protocol and the Challenges posed by Autonomous Weapon Systems

Rules of Engagement (ROE) are a positive statement of intent, underpinned by legal, policy, capability and operational factors that are specific to a particular theatre of operations. ROE are designed to ensure that the lethal force is employed in legally allowed manner. In practice, ROE have a two-fold role. First, they provide commanders at the operational and tactical level with greater control over the implementation of the military operation by their units. Second, they provide soldiers with clear guidelines on what is permissible on the battlefield. If we consider the assumptions of the security protocol theory in light of which a security protocol is defined as a codified system of rules and procedures that rationalize, legitimize and explain complex security forms, methods and measures and the procedures for their enforcement, it can be assumed that ROE, as a sui generis code of conduct, constitute a security protocol in the context of military operations. A security protocol is a relatively stable normative system that justifies the forms and methods of management, regulation and control, while ROE are the result of political goals, strategies and tactics as well as legal norms, therefore they take the form of detailed rules of conduct implemented in military operational procedures. Thus, it is a protocol in the form of a code (set of rules) that facilitates the application of law and enables the implementation of political and military goals during military operations. The paper aims to explain what significance ROE have on the modern battlefield, as well as to demonstrate to what extent they fit into the security protocol model and whether they have special features that go beyond this model. However, taking into account the rapidly advancing development of military technologies and the changing face of modern wars (so-called remote warfare), it should be expected that in the future autonomous machines will take over some of the tasks traditionally performed by soldiers. These machines are called "autonomous weapon systems" (AWS), and tentatively defined as "weapons systems that, after being activated by a human operator, can select and engage/attack targets without human intervention". Although existing fielded systems do not yet show advanced autonomy in the selection of targets, it is worth considering whether future "fully" autonomous weapon systems would need ROE separate from those issued for the human elements of the force they are part of. Should the answer be in the affirmative, a set of further questions arise, such as: will such systems be able to cope with ROE adopted for military operations? What kind of changes will be required to adapt the situation-oriented ROE to automated decision-making processes on which both existing and potential future-machine-learning based systems rely on? These questions are far from theoretical. Contemporary military operations take place in a very complex legal, military and political environment, and whether or not we can "teach" AWS to follow ROE will be just as crucial for their operational utility, as it will be for their compliance with the law of armed conflict.