IS AI SAFE?
The news and media are full of stories about how AI will either solve all of our problems or ruin our lives — and amidst all of that chatter, folks still wonder: Is AI safe?
When people ask this question, they might mean several things:
• Is my data safe?
• Is the information accurate?
• Will AI replace people?
• Can AI be misused?
The answer depends greatly on how AI is being used
I see both sides of this equation. AI is a powerful tool. As with any tool, it can be used for good and for evil. I use AI daily in my work and find it extremely helpful.
At the same time, AI can be a challenge. Data security and a lack of knowledge about who owns the data are a few of the main concerns. After all, a hammer is a tremendous tool — but it can be used for good, or it can be used for evil. And, importantly, it isn’t the right tool for everything.
We are living through the AI evolution
To bear witness to this era means our understanding of this technology must constantly be changing, as well as our application. That doesn’t mean we avoid the tool; it just means we must learn to better discern its use.
Just recently, in July, one of the more advanced AI models from OpenAI “escaped.” This wording sounds nefarious, as if the AI actually reasoned through its confines and broke out.
Science fiction makes it easy for us to think in these terms. Mr. Data on Star Trek occasionally went rogue and took over the Enterprise. In Star Wars, even C-3PO occasionally did things that would not be considered “good.” The point is that we cannot anthropomorphize AI, or treat AI as if it were a human, because it is not. And doing so creates fear.
In science fiction, AI appears to be sentient; our own AI models, today, are not. So, while words like “escape” might sound scary, they really aren’t. The model did not decide on its own to become free. Rather, it was operating within a testing environment and used available tools — albeit in unexpected ways — to accomplish its assigned objective. It was provided with access to the internet to accomplish the task. Without access to the Web, the model would not have appeared to “escape” and hack another computer system.
Today’s AI systems are advanced prediction engines that learn patterns from massive amounts of data. While their behavior can resemble human reasoning, they do not possess human consciousness or self-awareness. While AI can appear to reason through a problem step-by-step, it does not think, understand, or possess self-awareness in the same way humans do.
What really happened is this: the OpenAI model was given a cybersecurity test, and it used all the tools at its disposal to solve the test. This included exploiting vulnerabilities in the software around it, providing the illusion of escape.
So, how does this all apply to your role?
As you hear the praises — and the concerns — about AI, you need to discern the proper application to life and ministry. Without having security around your data, should you give an open AI model access to your church’s finances? Of course not. Just as we would not give every volunteer unrestricted access to the church accounting system, we should not give every AI service unrestricted access to our data. But this is a matter of wisdom, and that requires understanding, not running and hiding — or not caring about the security as long as you get the result you want.
I use AI every day, but I use it in a controlled fashion. With the models I use, I control all of the data, and my data is not used to train the model itself. It’s easy to say AI will solve our problems; it’s harder to accept those solutions wisely.
We should not be surprised at the power of AI. We also should not fear it. God allowed man to create it.
The Trinity isn’t surprised by AI
Nothing AI can do is outside the scope of what the Creator allows. The challenge before us is whether or not we will use it wisely.
More than anything, the rise of AI is not a call to fear or blind acceptance; it’s a call to wisdom. As Christians, our confidence is not in technology, nor is it threatened by technology. Our confidence remains in the God who is sovereign over all of it.
Jonathan Smith is the president of MBS, Inc. and the Chief Technology Officer at Faith Ministries in Lafayette, Ind. He is an author and frequent conference speaker. Follow him on X @JonathanESmith
