At first glance, the combination between AI and theology may sound like strange bedfellows. After all, what does technology have to do with spirituality? In our compartmentalization-prone western view, these disciplines are dealt with separately. Hence the first step on this journey is to reject this separation, aiming instead to hold these different perspectives in view simultaneously. Doing so fosters a new avenue for knowledge creation. Let’s begin by examining an AI worldview
What is AI?
AI is not simply a technology defined by algorithms that create outcomes out of binary code. Instead, AI brings with it a unique perspective on reality. For AI, in its present form, to exist there must be first algorithms, data, and adequate hardware. The first one came on the scene in the 1950s while the other two became a reality mostly in the last two decades. This may partially explain why we have been hearing about AI for a long time while only now it is actually impacting our lives on a large scale.
The algorithm in its basic form consists of a set of instructions to perform, such as to transform input into output. This can be as simple as taking the inputs (2,3), passing through an instruction (add them), and getting an output (5). If you ever made that calculation in your head, congratulations: you have used an algorithm. It is logical, linear, and repeatable. This is what gives it “machine” quality. It is an automated process to create outputs.
Data + Algorithms + Hardware = AI
Data is the very fuel of AI in its dominant form today. Without it, nothing would be accomplished. This is what differentiates programming from AI (machine learning). The first depends on a human person to imagine, direct and define the outcomes of an input. Machine learning is an automated process that takes data and transforms it into the desired outcome. It is learning because, although the algorithm is repeatable, the variability in the data makes the outcome unique and at times hard to predict. It involves risk but it also yields new knowledge. The best that human operators can do is to monitor the inputs and outputs while the machine “learns” from new data.
Data is digitized information so that it can be processed by algorithms. Human brains operate in an analog perspective, taking information from the world and processes them through neural pulses. Digital computers need information to be first translated into binary code before they can “understand” it. As growing chunks of our reality are digitized, the more the machines can learn.
All of this takes energy to take shape. If data is like the soul, algorithms like the mind, then hardware is like the body. It was only in the last few decades when, through fast advancement, it was possible to apply AI algorithms to the commensurate amount of data needed for them to work properly. The growth in computing power is one of the most underrated wonders of our time. This revolution is the engine that allowed algorithms to process and make sense of the staggering amount of data we now produce. The combination of the three made possible the emergence of an AI ecosystem of knowledge creation. Not only that but the beginning of an AI worldview.
Seeing the World Through Robotic Eyes
How can AI be a worldview? How does it differ from existing human-created perspectives? It is so because its peculiar process of information processing in effect crafts a new vision of the world. This complex machine-created perspective has some unique traits worth mentioning. It is backward-looking but only to recent history. While we have a wealth of data nowadays, our record still does not go back for more than 20-30 years. This is important because it means it will bias the recent past and the present as it looks into the future.
Furthermore, an AI worldview while recent past-based is quite sensitive to emerging shifts. In fact, algorithms can detect variations much faster than humans. That is an important trait in providing decision-makers with timely warnings of trouble or opportunities ahead. In that sense, if foreseeing a world that is about to arrive. A reality that is here but not yet. Let the theologians understand.
It is inherently evidence-based. That it is, it approaches data with no presuppositions. Instead, especially at the beginning of a training process, it looks at from the equivalent of a child’s eyes. This is both an asset and a liability. This open view of the world enables it to discover new insights that would otherwise pass unnoticed to human brains that rely on assumptions to process information. It is also a liability because it can mistake an ordinary even for extraordinary simply because it is the first time it encounters it. In short, it is prime for forging innovation as long as it is accompanied by human wisdom.
Rationality Devoid of Ethics
Finally, and this is its more controversial trait, It approaches the world with no moral compass. It applies logic devoid of emotion and makes decisions without the benefit of high-level thinking. This makes it superior to human capacity in narrow tasks. However, it is utterly inadequate for making value judgments.
It is true that with the development of AGI (artificial general intelligence), it may acquire capabilities more like human wisdom than it is today. However, since machine learning (narrow AI) is the type of technology mostly present in our current world, it is fair to say that AI is intelligent but not wise, fast but not discerning, and accurate but not righteous.
This concludes the first part of this series of blogs. In the next blog, I’ll define the other side of this integration: theology. Just like AI, theology requires some preliminary definitions before we can pursue integration.
