Implementation strategy

In the history of multilateral disarmament, African states are more effective when they act as a block. The chances of a lead-state emerging from Africa are more likely when the regional block – the African Union and its organs – have taken a position on killer robots.

1. Implementation strategy

The strategy is to first get a strong position by the African Union bodies, namely, the Pan African Parliament [PAP] and the African Commission for Human and Peoples' Rights [ACHR].

a) Course of action 1 – Representations to PAP

There should be efforts to seek representations on killer robots in the upcoming sittings of PAP Permanent Committees sessions and Ordinary sessions in Midrand, South Africa.

The aim of the representations is to have a PAP resolution calling for a ban on killer robots, something similar to the 2018 European Parliament Resolution on killer robots. PAP, as the legislative body of the African Union, can also place the issue on the agenda of the African Union.
Prior discussions with PAP officials revealed willingness to cooperate on the issue of killer robots and it fits in the 2020 agenda of the African Union, namely, ‘Silencing the Guns’.

b) Course of action 2: – Reach out to PAP parliamentarians

Engaging PAP should also include liaising with the six South African PAP.

c) Course of Action 3: – Engage the African Commission for Human and Peoples’ Rights [ACHR]

It is important to engage the ACHR especially on the human rights implications of killer robots. Prior engagements with officials of ACHR have shown willingness to cooperate on issues of human rights and technology, including the issue of killer robots.

d) Course of action 4: – Finding a lead-state

Once there is a strong regional position on killer robots, South Africa has a potential to be a lead-state on killer robots. South Africa is preferred because it is currently active in the killer robots discussions in the UN and cooperates with civil society in Africa.

2. Obstricals & solutions

First, since African States are not developing killer robots and many have no capacity to do so, there may be questions as to the relevance and urgency of the killer robots issue to Africa.

To overcome this, when engaging African States, it is important to emphasise that in the UN debate on killer robots, the voice of African States is critical, in particular, on issues relating to human dignity and the implications of killer robots on racial and gender discrimination. Furthermore, killer robots may be used against African States and their citizens.

Second, the issue of killer robots is highly technical and specialised; some African States may not have the technical capacity to participate in the killer robots debate. There may be limited accessible experts on killer robots from the African continent.

This can be overcome by offering technical support to relevant AU organs like PAP, ACHR and the African Group on Disarmament.

Political will exists to realise this proposal

Prior engagement with African regional organs such as the Pan African Parliament [PAP] and the African Commission on Human and Peoples’ Rights [ACHR] on the issue of killer robots have shown that there exists political willingness to engage on the issue of killer robots.

For example, engagement with the African Group on Disarmament in Geneva in 2017 and 2018 showed political willingness on the part of African States which was coupled with improved participation in the UN Group of Governmental Experts on Lethal Autonomous Weapon Systems (UNGGE).

The African Group of Disarmament submitted two statements to the UNGGE calling for the ban of killer robots. In emphasising that killer robots should be banned before they come into existence, the African Group of States noted as follows: “Notions of human dignity and humanity are the parents of all other humans’ rights and should govern human conduct including human inventions. The African Group finds it inhumane, abhorrent, repugnant, and against public conscience for humans to give up control to machines, allowing machines to decide who lives or dies, how many lives and whose life is acceptable as collateral damage when force is used. The African Group strongly asserts that technology should be solely dedicated for the prosperity and progress of human beings in all spheres of life.”

Furthermore, in its General Comment Number 3 on the Right to Life, the African Commission on Human and Peoples' Rights stated that “in any machine autonomy in the selection of human targets or the use of force should be subject to meaningful human control”.

Moreover, of the 30 states that have so far explicitly called for a ban on killer robots, 8 are African states. Against this backdrop, the proposal to develop a strong regional position is feasible and within reach.

What if political will does not exist yet

As indicated above, once the regional organs take a strong position on killer robots, African States will be able to participate more decisively on the matter of killer robots

Mitigating the threat from new and emerging technology

Killer robots raise far-reaching legal, ethical and security concerns. They cannot comply with laws applicable to armed conflict and will undermine protections afforded to civilians. They will commit war crimes and yet it will be difficult if not impossible to hold someone accountable. Indeed, it is unethical for machines to make decisions about who lives or dies. Killer robots will make it too easy for States to use force, thereby undermining international peace and security. Further, killer robots present capabilities that may undermine the survivability of nuclear arsenals and potentially render them vulnerable resulting in overwhelming threats to global security. There is an URGENT need for a legally binding instrument outlawing killer robots.

Yet, with the current threats to multilateralism which have seen powerful states walking back on their commitments on disarmament issues, it is a tough battle to produce a ban on killer robots in time, especially within consensus-based institutions like the UNCCW. This is made worse by the AI arms race between global powers such as the United States and China which may lead them to deploy killer robots "without adequate safeguards against brittleness, unpredictability, and systemic risks that will arise when algorithms interact at machine speed.”

Hope in the mitigation of this global catastrophic risk may rest with majoritarian-based institutions like the UN General Assembly or other processes outside the UNCCW. For this to succeed, there is need for like-minded States and regional blocks to come together and ban killer robots.

Getting the African regional block to take a strong position on killer robots and finding a lead-state from the African continent will be a huge step towards provision of alternative process outside the UNCCW. As already indicated, it is easier for African States to take positions if they are backed by the regional bodies.

Reducing inclusivity and accountability in national and global governance

A legally binding instrument prohibiting the development and use of killer robots will, in fact, reduce chances of algorithmic bias in the use of force. It will also that humans remain accountable for the use of force both domestically and at the international level.

Killer robots create an accountability gap. Because they are unpredictable, it would be difficult or impossible to establish legal responsibility of humans in cases where killer robots commit crimes. In such cases, killer robots undermine accountability, a very important component of global governance.

Further, in 2019, UN Group of Governmental Experts on Killer Robots and the UN Working Group on the Rights of Peoples of African Descent, noted that the development and potential use of killer robots may worsen racial and gender discrimination which can be perpetuated through algorithmic bias.

As such, this proposal is in the interest of inclusive society that holds criminals to account.

Conversing effect in increasing poverty and inequality

One of the questions that is asked is whether a legally binding instrument banning killer robots will negatively affect other AI technologies meant for civilian use like those that can be useful in development and poverty alleviation.

On several occasions and in this proposal, it is made categorically clear that a ban on killer robots will not negatively impact on the civilian uses of AI technology. If anything, with a ban on killer robots, States and other stakeholders will focus and invest more on the research and development of AI technology that is meant for the good of humanity instead of death and distraction.

Just as the current theme of the African Union states “Silencing the Guns: Creating Conducive Conditions for Africa’s Development”, a ban on killer robots will create a conducive environment within which AI applications meant for development and poverty alleviation are given a priority.

Reducing conflict and political violence

As already indicated above, there is already an AI arms race between major powers that threaten global security. In January 2020, the United Nations Secretary General, Antonio Guterres, identified killer robots as one of the four main threats to global security. There are scholars who have identified killer robots as one of the likely causes of World War III.
Killer robots will make it too easy for states to go to war and will reduce the threshold at which force is used in law enforcement situations. Just as armed drones were once thought to be the only game in the fight against terrorism, they ended up marking unprecedented deaths of civilians, among them, women and children.

Furthermore, killer robots may be used for politically motivated violence, in particular, political assassinations. Targeted killings on the basis of facial recognition will become more prevalent. In contexts of demonstrations, for example in contested elections in Africa, it may make it too easy for dictators to crush any opposition.

Therefore, a legally binding instrument banning or prohibiting killer robots will go a long way in avoiding future conflicts and political violence.

Of course, there are some scholars who have argued or suggested that the use of killer robots will reduce or do away with the brutalities that are currently being committed by human soldiers. The main basis of this suggestion is that killer robots will not act out of prejudice or ill-will. Yet, this suggestion has already been disproved by the existence of algorithmic bias and prejudice against people of colour.

Additional information

For my published articles on killer robots, see my webpage here: https://www.southampton.ac.uk/politics/about/staff/tc1f19.page

Files attached

TOGETHER FIRST PROPOSAL.pdf