A unique feature of Anglican life in the United Kingdom is that 26 Church of England diocesan bishops sit in the House of Lords, the second (and unelected) chamber of Parliament. I had the honor of being one of them for 10 of my 15 years as Bishop of Coventry. Among the challenges and opportunities that came my way was being a member of the AI in Weapon Systems Select Committee, which was appointed to “consider the use of artificial intelligence in weapon systems.” The committee was made up of 14 members drawn from across the parties, with our eyes especially on the development of autonomous weapon systems. It was fascinating and frightening at the same time to hear testimony from experts in the government and the armed forces, industry, academia, NGOs, and pressure groups, especially when several pioneers were warning of the existential risks to humanity because of developments in AI.
The committee reported as required within a year of its commencement, with its position clear from the title of its report: “Proceed with Caution.” The report’s summary clearly said that the U.K. Government “must ensure ethics are at the center of its policy, including expanding the role of the Ministry of Defence’s AI Ethics Advisory Committee” (p. 4). Also critical, from my perspective, was the firm view spelled out in the fourth of its central recommendations: “The Government should ensure human control at all stages of an AWS’s lifestyle.”
There was much that I learned from those who gave evidence to the committee, orally or in writing, and through site visits to research centers and military establishments. I gained a level of technical knowledge and understanding of the development and use of AI across the military spectrum, much of it uncontroversial. I also found, though, that the moral position with which I joined the committee was reinforced through the course of its work. My focus from the beginning was on what might be called the distinctive dignity of humanity. Theologically, such a position in the Judeo-Christian tradition is rooted in the belief that human beings are made in the image of God, according to his likeness (see Gen. 1:26).
As those made in the image and likeness of God, human beings have the capacity to reason. We use our intelligence in all its different forms to come to particular judgments that, at their best, show wisdom. Created through God’s Word, we are endowed with reasoning capacities that can be lifted by his Spirit to reflect even the mind of Christ (Phil 2:5; 1 Cor. 2:16). Machines, however sophisticated and technically proficient, are very different. Whatever “intelligence” they have is not the same as the intelligence with which human beings are endowed. Indeed, it is not intelligence as we know it. It is computation through data analysis, algorithms and the like; and it may be better to call it what it is rather than imputing human characteristics to inanimate things, often only having a digital reality.
Anthropomorphizing machines and their calculations is not at all helpful. It can distract us from what they are and deceive us into thinking that they have the same capacities as we have, and better. Theologically, the stakes are high. Projecting the features of human beings who are made in the image and likeness of God onto machines, which are by no means as wondrously created (Psalm 139.14), is only one step away from idolatry. The stakes are high pragmatically as well, especially when we begin to assume that through the exercise of “intelligence,” weapon systems have the competence to make life-and-death judgments in relation to human beings made in the image and likeness of God. In an important study; An Ethical Evaluation of Lethal Functions in Autoregulative Weapons Systems to be published next year by Nicole Kunkel, drawing on the philosopher Brian Cantwell Smith, makes an important distinction between the “reckoning” of which artificial systems are capable and the “judgment” that human beings are enabled by God to exercise.
On the modern battlefield — and the battlefields that are sure to come — it is important to retain confidence in human judgment, relying as it does on seeing the whole. It has a God-given freedom “to do otherwise” than what it is programmed to do, and it has capacities for empathy, for mercy, for perception in depth and for wise discernment that may transcend that which may be calculated even by the normal exercise of logic.
Alongside the gift of reasoning comes the calling of responsibility. To act responsibly means to make judgments that bear moral scrutiny, and to which we can expect to be held to account. They are an exercise of moral reasoning that follows from our creation in the image and likeness of God that embody in some way, albeit provisional and partial, the moral law of God by which we ourselves will be judged by God’s perfect justice. Acting ethically is more than reckoning whether our decisions meet certain programmed criteria. It requires those advanced attributes of understanding, the capacities of judgment I have already described that can see in the round, and the ability to assess short- and long-term consequences. All of these are needed for the responsible analysis of risk, upon which the humane conduct of warfare depends.
It is for these reasons that retaining proper levels of human control — meaningful human control, as it is often called — is essential in weapon systems. Quite what that means in relation to every weapons system in all its different elements, from design to targeting and different theatres of usage, is a complex matter in practice. But the principle is vital to uphold. The necessity of determining who is responsible for lethal decisions is necessary as much for law as it is for ethics. It is necessary theologically as well. For the calling to act responsibly lies at the core of our human identity, made in the image and likeness of God. To evade that calling, and to pass it on to that which is not human, is a denial of humanity and will lead to the deformation of humanity.
There are larger questions to be asked of the use of AI in weapon systems and the development of so-called Autonomous Weapon Systems. They are well articulated in Nicole Kunkel’s forthcoming book, and they apply to all weapons and the way that they are used. Will they serve the purposes of peace? That is not only a question of whether they limit and control the use of violence in warfare in the way that Christian thinking about the just use of weapons has sought to do (jus in bello), which is enshrined in international humanitarian law. It is also a question of whether they will serve the establishment of peace, the question that lies at the root of the Christian just war tradition (jus ad bellum) that reaches back to Augustine in the fifth century and, moreover, whether they will help or hinder the creation of a just and sustainable peace when the conflict is finally over in the way that more recent reworking of the just war tradition has done (jus post bellum).
I will soon lead the funeral of a distinguished old soldier. His family has asked for the Collect of the Argyll and Sutherland Highlanders, his first regiment, to be prayed at the service. The collect concludes with these words: “enable us, while loving our country best, to enter into the fellowship of the whole human family, and give us now and ever the gift of courage to seek after a just and merciful peace.” I shall offer that prayer, remembering not only the life of a fine human being who served his country well, but also praying for those charged with the responsibility of the development and use of weapon systems, especially those that incorporate Artificial “Intelligence.”
