Group of Governmental Experts on Lethal Autonomous Weapons Systems – First Meeting – Item 5(B)

Group of Governmental Experts on Lethal Autonomous Weapons Systems

First meeting: 25–29 March 2019

STATEMENT BY ESTONIA

 

Agenda item 5(b). Characterization of the systems under consideration in order to

promote a common understanding on concepts and characteristics relevant to the

objectives and purposes of the Convention

Thank you, Mr Chair!

Permit me to begin by congratulation you on your election as Chair of the GGE. The

Estonian delegation appreciates the practical and detail-oriented approach that you have

taken in organising our work; we have no doubt that under your able guidance we will be

making significant progress.

Estonia aligns itself with the statement delivered by the European Union, but would like

to make some observations in a national capacity. In response to your invitation to

proceed in a practical manner, we have a few comments directly in response to the

questions that you have posed for us.

First, we see autonomy as a function-specific property.

A ‘weapon system’ – the term used in the name and mandate of this GGE – generally

refers to a weapon along with any associated technology necessary for the operation of

the weapon. Different component parts of such a system necessarily have different

functions or tasks. If we extend our enquiry to weapon platforms – the structures and

systems onto which weapons can be mounted – the range of functions and tasks increases

further.

In different weapon systems and platforms, these functions or tasks may be subject to

different degrees and types of human control. For example, an unmanned aerial combat

vehicle might have significant autonomy in movement (including take-off, navigation and

landing) but no autonomy in targeting.

In our view, therefore, autonomy relates to specific functions or tasks of weapon

systems and platforms, and autonomy can have different degrees with respect to each

of those functions or tasks. As such, we think that the notion of a fully autonomous

weapon system has limited utility.

Second, IHL is potentially important to all functions of weapon systems.

PERMANENT MISSION OF ESTONIA

TO THE UN AND OTHER INTERNATIONAL

ORGANISATIONS IN GENEVA

The relevance of IHL to functions involved in target selection and engagement is obvious

because IHL extensively regulates attacks.

But we note here Article 57(1) of Additional Protocol I to the Geneva Conventions. This

provision requires states to take constant care to spare the civilian population, civilians

and civilian objects not just in attacks but in the conduct of military operations more

generally. This suggests that other functions, such as the autonomous movement of a

weapons platform, need to be developed and implemented with the safety of civilians and

civilian object in mind.

That does not mean, however, that the GGE needs to consider the totality of IHL rules in

relation to any autonomous functions of weapon systems or platforms. The GGE operates

within the framework of the Convention on Certain Conventional Weapons. As such it

seems entirely appropriate to focus on the increased autonomy in the critical functions of

a weapon system, that is selecting and engaging targets. As such, platforms that can move

autonomously but rely on a human operator for targeting decisions, properly fall outside

the scope of our discussion.

Second, we consider the environment where a weapon system is deployed to be highly

significant.

To some degree this significance arises from existing international law, which varies

slightly based on the domain of warfare. For example, landmines and naval mines are

addressed by different treaty-based rules.

But even where the applicable rules are the same, the degree and nature of human

oversight required to ensure compliance with IHL depends on the operational

environment. For example, the human control to be exercised over weapon systems in

urban warfare, where the risk of harming civilians is particularly acute, may well be

different from the control required in submarine warfare, where risks to civilians are more

limited.

Attempts in this forum to determine the requisite degree of human control over weapon

systems should be mindful of such legal and operational differences.

Finally, we consider the distinction between anti-personnel and anti-material weapons to

be meaningful in the context of our discussions.

Certainly, anti-materiel weapons, irrespective of any autonomous functionality, can also

cause injury to human beings. Also, IHL rules plainly apply to all weapon systems used

in armed conflict, irrespective of their intended use.

But humanitarian concerns are heightened with respect to certain types of weapons

designed for use against human beings. This is clearly reflected in the focus of some of

the protocols of the CCW and other international law instruments exclusively on antipersonnel

weapons. Also, the existing weapon systems with autonomous functionality

that have been mentioned in this forum are essentially anti-materiel weapons. The

uncertainties about addressing legacy systems with autonomous functionality are

therefore largely confined to anti-materiel weapons.

I thank you, Mr Chair!