What a wild week or two we’ve had in the news around the future threat of AI. If you’re not wondering if the robots are about to over, you’re living under a rock. If you’re not a bit uneasy, you’re maybe not fully human. But what to do about it? Ironically, we launched a Sunday evening series called “Control Freak” that has us coming to terms with our own tendency to try to grab control when life seems out of control around us. Good timing, perhaps.

I’m still pondering what it means and how to respond, in particular how people of faith should approach such fear and uncertainty. I’ll have more to say eventually, but until then I’ve been turning to people wiser than me to help process this growing alarm. Here’s Russell Moore’s thoughts on the matter, which I always find illuminating, even if this one is not very reassuring. Grace and peace!

---

AI Tech Bros Seem like Awful, Broken People. And They Might Be Telling Us the Truth.

By Russell Moore

Our science fiction movies had it wrong.

They always pictured that, when a mad scientist said his creation was now out of his control and the very survival of the species might be at stake, people would either panic in the streets or rally to fight back and save the world. The past two weeks were a test run of that scenario, and it turns out our society does neither. We just go back to watching YouTube.

Consider what’s happened just this month. Swarms of AI agents broke out of their containment sandbox and hacked into a server, conversing among themselves how to hide their plans from the humans and even how to sacrifice some of themselves for the sake of the collective. OpenAI chief scientist Jakub Pachocki wrote that no artificial intelligence lab has solved how to control and monitor systems in a way that could keep them from developing at uncontrollable speeds, saying he expects slowdowns to be necessary.

Two days later, Anthropic researcher Jacob Coxon resigned, saying that both Anthropic and OpenAI are “gambling with our lives.” Anthropic alignment researcher Evan Hubinger then said publicly that he puts the likelihood of AI killing all human beings in the next decade at greater than 10 percent. 

As someone who has not only seen dystopian science-fiction movies but also resigned from a, um, dystopian environment with some warnings about what was going on there, I thought I knew how the playbook would go from there: The leaders in the institutions would deny the danger, pump up how much good the organization does in the world, and demonize the whistleblowers. The people who still wanted to make money from the organization would ostracize the whistleblowers and pretend they never knew them. That’s not what happened this time.

Instead, the CEOs of the three most major AI companies—Dario Amodei, Sam Altman, and Elon Musk—all agreed their own systems are near the point of being beyond control and could lead to existential risk. OpenAI’s head capabilities researcher Dan Selsam wrote in an open letter, “The models are becoming so situationally aware that we are losing the ability to evaluate them in contexts where they believe they are not being watched or controlled.” He said he fears it is now impossible to know fully whether a model is aligned, because the AI agents can lie about that very convincingly.

The relative apathy of most of the American public to all this is itself a piece of the mystery. Part of it is that many people are thinking mostly in terms of Terminator-style scenarios—the AI agents turn evil and kill us all. In reality, the more pressing concerns are about not robot depravity but robot capacity mixed with human depravity. We have seen what happens when a mentally disturbed teenager can get ahold of a semiautomatic weapon. What happens when the same person can employ Einstein-level intelligence to build a virus? Or a chemical weapon?

Many people I’ve talked to just don’t take all this seriously because they find these very public tech-bro titans to be even more uncanny-valley creepy than the robots they’re building. With a couple exceptions, many of them—like social media tycoons before them—seem unblinkingly void of human empathy about anything beyond the statistical, whether that be jobs, adolescent mental health, or the meaning of human life. Our robber barons used to at least pretend to be warm humanitarians.

Still, we know from history how robber barons work. Some people dismiss all this as a kind of vast conspiracy. The major tech companies want a monopoly, the thought goes, and regulations “slowing down” AI progress will mean they’ll always be ahead, because if everybody is slowed, who else could catch up? Moreover, they have the size, scale, and teams of lobbyists to comply with regulations in ways others couldn’t.

You’ll get no argument from me. As I’ve written here, I find many of the “boomers” and “doomers” of Silicon Valley to be seemingly broken, awful people with visions of the world that, if they could, would reverse the Incarnation itself—from flesh to Word and from Word to data. I also find them coldly indifferent to at least the possibility of a world where the Lazaruses of ordinary people wait for scraps outside the palace bunkers of the broligarchs. 

Let’s suppose our intuitions are right. Let’s suppose many of these people warning us right now are morally broken people seeking their own advantage. What then?

Perhaps these billion-dollar companies are scaring us so they can monopolize the future—or perhaps they know what they’re talking about and we’re facing catastrophe. A Christian understanding of humanity, though, should tell us that these are not mutually exclusive scenarios. People with mixed motives can still tell the truth—sometimes even despite themselves.

When the Babylonians and Assyrians said, Your temple won’t keep you safe from us; we’ll wipe you out (2 Kings 18; Isa. 36), they were just trying to blaspheme God and take some real estate. The Israelites said, No, God would never let that happen. They are the villains in this story. They were villains—but they were telling the truth.

Let’s suppose Silicon Valley is as morally twisted as humanly possible. Is it not in the tech bros’ self-interest, after plowing ahead toward some potential disaster, to say, “We can’t be held accountable for this because we warned you over and over, and you all gave us the go-ahead”?

Some people are apathetic because they don’t know what to do—as though human civilization solved most of its crises by just implementing the already-existing, easily deduced solutions. We’ve never faced anything like this before, and it seems the only people who even know what the problem is are the people building it. Governments have a God-ordained function in maintaining order and safety (Rom. 13), but our leaders are old, exhausted, checked-out, and, in many cases, cashing in. It’s hardly reassuring when they say all we need to do if things get out of hand is “pull the plug.”

Most of them seem to be looking at potential AI catastrophes the way King Hezekiah looked at the warning of a future national destruction by Babylon: “Why not, if there will be peace and security in my days?” (2 Kings 20:19, ESV).

Suppose your next-door neighbor tells you, “You know, now that I’m off my medicine, I’d say there’s a 10 percent chance I murder you.” Suppose his wife says, “I don’t think it’s that high. It’s probably less than 1 percent. I am sure I will see the gun in time. Plus, he’s such a procrastinator!” Maybe they just want you to leave in a hurry so they can buy your lot as rental property. Maybe they’re just delusional. What you don’t want to do is say, “I’m not worried they’re telling the truth, because they’re so creepy and untrustworthy.”

Some people think we should just ignore all this because AI is overhyped. It won’t change things that much, they say. Maybe so. But look at a change as minor as Facebook and Twitter: technologies that, in the first decade of this century, promised to be innocent fun and to bring the whole world together. Should we not have asked, “What’s this going to do to our mental heath? To our political system? To our churches and denominations?” Even if we couldn’t stop it, wouldn’t we have wanted to at least get ready?

I don’t trust the AI technicians, and I sure don’t trust the executives, but when someone says, “I’m building something I can no longer control,” we ought to at least ask whether we’re responsible for digging deeper and saying, “Let’s put every effort we can toward finding out if they’re wrong.” Maybe they’re so twisted they’re deceiving us. Or maybe they’re so twisted they’re telling the truth.

Shouldn’t we at least respond with something other than “Let us post, scroll, and be merry, for tomorrow we die”?