Everyone agrees that Artificial Intelligence technology (AI) will bring profound changes to civilization, but no one is sure exactly how much change — or whether that change will be good or bad.
I first dug into this issue two years ago when ChatGPT exploded into the news due to its uncanny ability to converse like a real human. Previously, I had thought that computers would never equal human intellect, and this development rocked my assumptions about technology. Was humanity about to be eliminated by HAL 9000s, or was the hype much ado about nothing?
Nervously, I read more about AI, delving into the nuts and bolts of what was actually happening inside those microprocessors. What I found was both comforting and concerning. AI is not, and I believe cannot ever be, truly conscious. Nevertheless, the technology poses physical and spiritual dangers to society.
In this six-part series, I hope to make AI more understandable, which alleviates fear of the unknown. I’ll also talk about why the robot apocalypse is unlikely to occur. Then I want to point out some of the unique hazards of AI from a Christian perspective, concluding with some thoughts on how Christians can AI-proof their own lives and the lives of their children.
What Is AI, Anyway?
First, and most importantly, the moniker “Artificial Intelligence” is extremely misleading. There is nothing “artificial” about AI programs; they are highly engineered algorithms with an enormous amount of design. AI programs also rely on human training. In other words, they must be fed reams upon reams of human-crafted articles, photographs, and other data before becoming usable.
The “intelligence” part is also misleading. As we’ll discuss later in this series, there is no mind in an AI program. There is nothing that truly thinks or rationally weighs options — and under the right circumstances, AI can even fabricate data or advocate for suicide. The utility of AI follows the old programmer’s maxim: Garbage In, Garbage Out.
The true miracle of AI lies in pattern recognition. Previously, only humans could assess large, imperfect data sets, discern a trend or pattern, then extrapolate from it to make something useful. This ability undergirds the brain’s process of learning a language, picking out a picture of a cat from those of other animals, or even designing a novel medicine based on existing molecules.
AI programs rely on so-called “neural networks,” which are thought to mimic how the human brain functions. Imagine a very large grid standing on end. Each square holds an equation that can put out a word, and a weighting factor determines how often that word will appear. The neural network is trained by running existing text through the equations from top to bottom. As more text is run through the program, the weighting factors are adjusted so that some words become more likely to appear when found next to other words. When the AI program is done with training, a user can ask a question to generate text based on those stored probabilities. Image-based AI programs work differently than this, but still use a neural network.
Thus, a more useful term to which one might apply “AI” would be “Algorithmic Interpolation.” The programs are not generating anything completely novel, per se, but instead put out something that “looks like” whatever has previously been fed through them. Or more accurately, the output resembles data that has been stored in the weighting factors of a huge number of equations in the virtual grid. With this ability to create something that “looks like” preexisting things, AI can write an entire essay based on a simple prompt, mimic someone’s voice, make an image that a skilled graphic designer would be proud of, or even generate entire movies starring long-dead actors.
Consequently, AI has great potential to free people from the most repetitive, dull tasks on which our technology-driven society relies. It may also speed up discoveries in materials science or medicine due to its ability to analyze large data sets. Using AI, laymen can unlock creative impulses that formerly would have required years of schooling in computer science or a whole movie production team. Truthfully, no one is quite certain about all the eventual uses of AI, but industry, governments, and the stock market are betting heavily that it will provide them with a competitive advantage.
Every tool, though, no matter how useful, also can be abused. In Part 2, I’ll lay out some of the dangers presented by AI – both to our physical world and our spiritual one.
Brandon Aldinger is a chemist with a doctoral degree who works in an industrial research laboratory. He’s had lifelong interest in issues of science and faith, and he is passionate about training fellow Christians to think clearly about and stand firm on their beliefs within a hostile culture.
