Ethical Guidance on AI
Pope Leo XIV issued his first encyclical – an authoritative teaching on a specific moral question that takes the form of a letter – last month. It was subtitled “On Safeguarding the Human Person in the Time of Artificial Intelligence” but it has a much wider remit, offering extended and careful reflection on a range of topics from the history of Catholic Social doctrine to the legacy of the Atlantic slave trade to the relevance of “Just War theory”.
Pope Leo XIV is very self consciously positioning himself as a successor to Pope Leo XIII, who was a Pope in the late 19th century renowned for the groundbreaking encyclical that is considered the starting line for Catholic Social doctrine. In Rerum Novarum, the earlier Pope addresses the “new things” that emerged with the industrial revolution.
In that era of untrammelled capitalism, Leo XIII argued that the owners of factories had obligations to their staff and to the common good. The workers had a right to organise and should expect a liveable wage. Everyone, he insisted – worker and owner – could flourish together. “Capital cannot do without labor, nor labor without capital. Mutual agreement results in the beauty of good order.”
What’s to be Done
We face new things again, with the emergence of the digital revolution. Our societies have been transformed over the last generation by the acceleration in personal computing power, the expansion of the internet, and the proliferation of smart devices. Pope Francis described what it feels like to live in this cultural moment as “rapidification”. And it seems things are only getting more and more rapid.
This is the reality that Leo is trying to address in Magnifca Humanitas, and he is bold enough to offer some suggestions as to how we can flourish in the midst of ubiquitous AI and the ever-watchful gaze of the algorithms that track our every move, every spend, and almost our every word.
Where Pope Leo XIII had to argue for seemingly unrealistic goals, like time off every week, living wages, and safe working conditions, Leo XIV argues that we need clear regulatory oversight of algorithmic and AI technologies so that “responsibility must be clearly defined at every stage.” When you hear about Large Language Models encouraging people to end their own lives, you start to see in the starkest possible terms how important it is that the people who make, sustain, and profit from these technologies are held to account for the consequences of their deployment.
Because our lives are increasingly conducted online, we need to have a sort of preferential option for the technologically poor. No one should be left behind because they can’t afford to keep up. This applies to citizens in your nation and it applies to other nations. Vast amounts of data is being amassed by these tech giants, data that is personally linked to us. Data forged from your actions should ultimately belong to you. We must restore “to individuals not only the data that describes them, but also the ability to decide how it is used, by whom and for whose benefit. Otherwise, the digital age will not be post-colonial, but colonial in another form.”
And fundamentally, the Pope repeatedly insists that we must “disarm” AI. He means this in a metaphorical way, dealing with the power that gathers in the hands of the people who control these technologies and in the sense that they can deform our understanding of the truth. Leo writes:
“In the digital age, a just social order guarantees everyone equal access to opportunities, protects the youngest and weakest members of society, combats hate and misinformation and subjects the use of data and technology to public oversight, so that the guiding principle is not solely profit but the dignity of every person and the common good of all people.”
But he also means this quite literally. Since 2013, ethicists across the world have run a campaign called Stop Killer Robots. What was once considered a sort of speculative, sci-fi-ish over-reaction is increasingly a pressing reality. We know that militaries have already begun deploying autonomous systems with lethal effect but even the producers of these technologies are in the dark about the extent of their use.
For millennia, Just War theory has been one of the only tools at our disposal to hinder the temptation of leaders to wage endless war. Leo admits that this framework is no longer relevant. The weapon systems that are deployed in combat now make even the most basic guidelines like “Don’t hurt civilians” impossible to fulfil. If we – ordinary people – do not force serious and binding legal restrictions on the use of these new technologies, they will become embedded, accepted, and permanent features of our lives. Leo could not state it more plainly: “It is not permissible to entrust lethal or otherwise irreversible decisions to artificial systems.”
A Historical Reflection
Magnifica Humanitas formally echoes Rerum Novarum, but its content has a deeper historical resonance. That first social encyclical addressed the industrial revolution, which was sparked by a technological revolution brought about by steam power. This prospective new industrial revolution is similarly catalysed by a technological advance.
But as the Swedish geographer, Andreas Malm, has compellingly argued, steam power was not advanced because it was inherently more efficient or cheaper than the pre-existing technology. Indeed, when British manufacturers began to shift to coal-powered factories 200 years ago, water mills were more abundant, more reliable, and much less costly.
The attraction of steam power was not (initially) its increased mechanical power but its mobility. Running your concern on water meant you were tethered to a particular place. Fast-flowing rivers tend to be rural, and rural places tend to not have abundant labour and so the workers held significant bargaining power. The shift to coal decoupled capital from geography. Now, you could build a factory where the people were and, with that dense population of potential staff, ensure your wage costs were kept low.
This historical nuance is important to recall as we consider the potential economic impact of AI. AI is tangibly less effective than humans at the vast majority of work it has been tasked with. Its huge costs are being structurally obscured as it spreads. But this clunky, hallucinating, environmentally rapacious technology is still considered the bright hope of our economies.
The story of technological inevitability, that Leo is directly challenging, functions as a threat to workers. Pretending that these tools make some workers obsolescent and algorithmically surveilling those that remain is an excellent strategy to break the collective power of works, to drive down wages, and to forge a more compliant citizenry. Finding a way to disarm this mode of AI might be the most pressing challenge we face.
