How Not to Be Scared of AI – Sarah Eekhoff Zylstra, Joel Jacob

The headlines are enough to rattle anybody.

“The AI Arms Race Is Changing Everything,” Time announces. “How Will AI Change the Job Market?” Forbes asks. Over on Yahoo Finance, Raoul Pal tells us AI could become “the biggest bubble of all time.”

From classroom papers to fighter jets to medical surgery, it seems as if artificial intelligence is everywhere. As Christians, we don’t want to run in fear—after all, God is sovereign over robots too. But neither do we want to be reckless or careless in how we approach it (Prov. 14:16).

The Gospel Coalition sat down with Joel Jacob, who is a product manager working at the convergence of edge computing & AI. We asked what AI is, where it’s going, and how Christians can think well about it.

What is AI?

From the 1950s to the 2000s, people wanted to see if they could make computers exhibit intelligence indistinguishable from humans. During this period, it was attempted solely by using algorithms—which is basically a rule or set of rules you tell a computer to follow.

In the early 2000s, people began using a new approach called “deep learning.” Instead of writing code that told a computer what to do, people shifted to creating systems that replicated God’s design of the human brain.

To do that, a lot of research went into understanding how the brain works, starting with how babies learn words, sounds, and shapes. A lot of that comes back to the reinforcement of neurons in your brain based on environmental training. For example, we say “mama” and “dada,” and as babies replicate those sounds, neurons in their brain get trained and strengthened. We call that learning.

It turns out we could recreate a blank slate of neurons, and then treat a computer like a baby and train it based on data. What we end up with is software described in the industry as a “neural net” that is much closer to human intelligence than ever before—but with a speed advantage over humans since electricity on a circuit board is more than 100 times faster than our electro-chemical-based brain.

God is the ultimate Creator, and we’re always subservient to him. But we are creating something that could eventually think “better” and faster than we can.

Should we be scared of AI?

I wouldn’t say scared, but we need to be thoughtful because it’s a powerful technology. It reminds me of when scientists at the Manhattan Project created the atomic bomb. When the first tests were done, they were shocked to realize what they’d created. It’s kind of the same with AI, and whoever is ahead in this new technology race will have a powerful edge over those who are behind.

That’s why the OpenAI organization (the creators of GPT) was first created. The premise was that we don’t want one singular person or country to have control over this incredible power because of how it could be misused. It should be open so we can all equally move into this next era. But the narrative later flipped to “This technology is too powerful for everyone to have, so now we’re going to be careful with what is made public.”

God has allowed us to do this so far. But there are implications if it isn’t done in the right construct or doesn’t have the right restrictions. When we first created a bridge, it inevitably fell down and people got hurt. Then we learned from that and got better at building bridges (although even recently bridges have collapsed and hurt people). In our first attempt at AI, we know we aren’t going to get it right. How do we prepare for the fallout?

This is already coming to light as we notice bias or flat-out incorrect answers from these machines. We always need to remember that an AI’s deep learning is shaped both by the people coding the model and by the data fed into it. Theologically, we understand that we’re imputing our sinful nature to the machine.

Christian developers need to be thinking, What’s the best way to develop this ethically, in a way that will benefit humanity and hopefully bring them closer to God?

Where does it go from here?

AI will be adapted. Conversational AI or “large language models” like the ones used in ChatGPT will continue to improve. ChatGPT 4 is coming out later this year, which has about 100 times more parameters in the neural networks than 3.5 (the version out now). It will make fewer mistakes. In other spheres, even where we thought AI could never be meaningful, the first AI-generated art won a contest in September, and the first AI-controlled fighter jet flew more than 17 hours in December. It will be increasingly valuable to have the skillset to know how to best work or leverage AI.

Jobs will shift, but not for the first time. When we first invented tractors, we displaced three-quarters of the agriculture industry. That level of displacement may happen here. The World Economic Forum estimates that while AI will replace humans in 85 million jobs, it’ll also create 97 million new ones.

We can anticipate substantial quality of life improvements. If a robot running on AI could do chores around the house, that would be a big help to older people who struggle to get around or to working parents trying to stay on top of the housework. Or if you could ask AI to explain things to you, that could further accelerate access to knowledge—like how the internet made information more accessible to those who didn’t have access to libraries or know how to search through swaths of books. Of course, all of these changes come with tradeoffs, but for many the benefits outweigh the consequences.

What isn’t clear is the ceiling on how capable AI can become. Technologists debate the possibility of the “singularity,” or when technological growth becomes self-perpetuating and unstoppable, and when that might happen.

How am I supposed to think about this as a Christian?

There are two important things to remember:

1. God can still use machines for his general benevolence and good.

2. AI may end up being helpful, but it shouldn’t become an idol. And it can’t replace humans. The crux comes down to its purpose, identity, and autonomy.

While AIs may end up being more “intelligent” than humans, that does not provide them a spirit. Our status as the most intelligent creatures on the planet may not be as valuable as we hold it, and that is something challenging to wrestle with. But even if AIs could worship God—and there’s a good chance they will, since even rocks can cry out (Luke 19:40)—that still does not give AIs the same relationship with God that humans have.

We should think deeply about how the world will look and how the gospel could be accelerated in a world of AI, the same way we considered how the internet could benefit gospel growth.

I’m likely going to spend another 20–30 years of my life working in tech. If God gave me that as my mission, I’m asking myself, What should I do? How do I create tech that will bring more people to God? How do I use AI to point people to Scripture?

It’s also important when creating technology like this to consider my heart posture. If I’m thinking, “Oh, I’m God,” then I’m similar to the people building the Tower of Babel.

But I want to constantly be asking, “How can this bring glory to God?” I want to reflect his image by showing his miraculous creation and pointing people to Christ.

Read More

The Gospel Coalition

Generated by Feedzy