Hostname: page-component-586b7cd67f-t8hqh Total loading time: 0 Render date: 2024-11-27T20:10:08.232Z Has data issue: false hasContentIssue false

Artificial Intelligence

Published online by Cambridge University Press:  01 January 2024

Rights & Permissions [Opens in a new window]

Abstract

Type
Comment
Copyright
Copyright © 2023 Provincial Council of the English Province of the Order of Preachers

I was recently speaking to someone who was about to embark on a PGCE course in religious education. He mentioned that if he had chosen to become a mathematics teacher, the government would have paid him a lump sum of £25,000 just to do the training. The PGCE student in religious education on the other hand, receives no such perks, but rather must pay out thousands of pounds in tuition fees and living expenses in order to complete the teacher training. This is despite there being a huge shortage of RE teachers.

But we live in a culture that prizes mathematical ways of thinking, whereas theological ways of thinking are often seen as irrelevant and outdated. But there are signs that some influential people in our culture are beginning to realize the serious challenges humanity will face if we only understand ourselves and the rest of reality in mathematical terms.

One such sign is the recent discussions over Artificial Intelligence (AI). Currently, it seems that not a day goes by without AI being mentioned somewhere in the news. As I write this, today's news is that one of the so-called ‘godfathers’ of AI and winner of the prestigious Turing award, Prof Yoshua Bengio, feels lost over his life's work, and that in hindsight, would have prioritized safety over usefulness had he realized the pace at which AI would evolve.

Mathematical principles are fundamental to AI, but we cannot rely on mathematical principles alone if the power of AI is to be harnessed safely. There is an emerging consensus that AI development needs to be regulated, but it is hard to see what good can be achieved by regulation if those doing the regulating have an impoverished understanding of human existence and of what constitutes human safety. Already, there is widespread talk that mass unemployment could result from AI, but if AI regulation is done badly so that AI development is restricted to a small and powerful clique, mass unemployment could become even more of a problem.

But when it comes to questions of human existence and human safety, there is great value in a theological perspective. In our own Christian theological tradition, we understand ourselves to be made in God's image. There is a mystery at the heart of human existence that reflects the mystery of God. God has bestowed on us a nature by which we can take responsibility for our actions; we have been granted stewardship over the natural world, yet we look beyond this world for our ultimate fulfillment. And so God in His dominion over creation wants us to share in His dominion; He doesn't want to rule over us as an autocrat.

Now this understanding of human existence and our relationship with God ties in with the Christian notion of subsidiarity where those powers and responsibilities that can be effectively performed at a local level should not be usurped by a more central authority. However, the principle of subsidiarity is often ignored in the AI industry. For instance, many people involved in the development of self-driving vehicles aspire to the goal of level 5 automation in which a vehicle can manage all roadway and environmental conditions without there being any need for human intervention. But this inevitably means that decisions about human safety that could have been made by people actually travelling in the vehicle are going to be made by engineers at the development stage of these vehicles, engineers who are not going to face serious injury or death if they make a bad decision.

Another way in which a Christian theological perspective is of relevance to AI is in regard to the nature of human understanding. According to the Thomistic tradition, human understanding requires an immaterial intellect that can abstract from the contingent aspects of physical reality and recognize what happens of necessity. Being able to perform this level of abstraction means we can recognize when the conclusion of an argument necessarily follows from the truth of its premises. From a Thomistic perspective, I wouldn't expect something that is entirely material to have the capacity to do this in every situation. And if you try asking the current version of ChatGPT, it does indeed struggle to evaluate the validity of arguments, especially valid arguments that are unsound. It remains to be seen whether future versions of ChatGPT will one day master this capacity. If they did, this would surely challenge traditional arguments for the immateriality of the human intellect and the subsistence of the human soul. But even if one remains rather skeptical of claims that AI is on the verge of surpassing human intelligence, AI can still produce very plausible sounding explanations for the false claims that it makes. An AI machine that spews out convincing falsehoods could be more dangerous to society than an AI machine that is a reliable source of truth, so perhaps there will be an important role for AI regulation in helping to mitigate this risk.

Much more could be said on the possible ways a Christian theological perspective could be of relevance to AI and how it should be regulated. If regulated wisely, many good things could come from AI. But if governments continue to follow policies that are detrimental to religious education and which overemphasize the value of mathematics and the technology that follows from it, then very few people with a theological perspective are going to be involved in key decisions relating to AI development, and as a result, unwise policies may be adopted that do little to mitigate the risks AI presents to our society.