On the Tuesday, June 23, 2026, episode of The Excerpt podcast: In “The God Test,” Robert Wright argues that the real challenge of AI isn’t just building smarter machines—it’s deciding whether we can evolve, morally and politically, fast enough to develop an AI that’s mutually beneficial. Author Robert Wright joins The Excerpt to talk about his new book.
Hit play on the player below to hear the podcast and follow along with the transcript beneath it. This transcript was automatically generated, and then edited for clarity in its current form. There may be some differences between the audio and the text.
Podcasts: True crime, in-depth interviews and more USA TODAY podcasts right here
Dana Taylor:
Artificial intelligence is often framed as a story essentially about a technological evolution. It’s about breakthroughs, capabilities, and competition. But what if the real story is something much bigger? What if AI represents a turning point, not just for economies or industries, but for human civilization itself? An inflection point that forces us to confront questions about power, purpose, and what kind of species we want to become. In The God Test, Robert Wright argues that the real challenge of AI isn’t just building smarter machines, it’s deciding whether we can evolve morally and politically fast enough to develop an AI that’s mutually beneficial. As these systems begin to mirror and amplify human intelligence, agency, and even our flaws, the stakes become existential. Humans have shown that we can evolve. The question is, will we?
Hello, and welcome to USA TODAY’s The Excerpt. I’m Dana Taylor. Today is Tuesday, June 23rd, 2026. Robert, thank you so much for joining me.
Robert Wright:
Thanks for having me.
Dana Taylor:
You framed the challenge of evolving AI as one of purpose, writing, “Is it possible that the future like Gods of the past is demanding that we move towards spiritual progress, toward a kind of enlightenment, that we become better versions of ourselves?” Robert, talk me through the thinking behind this framing.
Robert Wright:
Yeah. Well, first of all, I think, and I don’t think I’m completely alone here, that if we’re going to handle this technology well, if we’re going to get through the AI revolution in good shape without certain catastrophes unfolding, possibly existential disasters, we’re going to have to approach this as a true global community. It’s an inherently international issue. There are just too many things that can go wrong that cross national borders.
I mean, starting with an AI maybe helping somebody build a bioweapon or something, but there are a lot of examples where it becomes clear that national policy alone is not going to be enough to keep things under control. So we’re going to need to have more in the way of international governance, international rules than we’re accustomed to having. And I think getting to that point is going to require a kind of moral progress, you could say, because we’re going to have to build a more peaceful world. We’re going to have to have much less in the way of conflict, especially including military conflict. And that in turn is going to call for us all to get better at looking at things from the point of view of other groups of people, other nations, even nations we think of as adversaries, even enemies. We need to at least understand how the people there and the governments there are looking at things if we’re going to develop these cooperative relationships that I think in light of this technology are absolutely essential. And I consider that a kind of moral progress.
If you get better at looking at things from the point of view of rivals, adversaries and so on, it’s hard. It’s very hard because there are specific cognitive biases that are built into people that make it hard, but I think it’s doable and I think it’s going to have to happen. This is a very powerful technology, its power is starting to become clear to more and more people. And I think in a way it’s good news that it almost forces us to get better at dealing with one another on a planetary scale.
Dana Taylor:
Rob, you place AI in a billion year evolutionary story, not just as a tech trend. Why is that perspective so essential?
Robert Wright:
Well, life’s been around for more than three billion years. This is the first time a wholly new form of intelligence, not even based in biological material, in other words, it’s not carbon based, has materialized. So, there’s that. That’s new. And as it happens, this coincides with the development, the kind of emergence of what you could call a global brain. I talk in the book about this thinker who in 1923 coined the term, “Noosphere,” for this kind of planetary mind that he could already see was emerging, and now with the internet, we can see more clearly.
So we’re going through these, a couple of big thresholds are being crossed. Information processing is reaching the planetary level. And now, I mean, that was happening anyway. 30, 40 years ago, connections among human brains were forming this brain of brains as this thinker, Teilhard de Chardin, had put it. But now we have this wholly new form of silicon-based intelligence that itself will be, these will be neurons in this global brain. So you got these two things happening at once, and they are both epic even against the backdrop of all of life. So, that’s what I mean when I talk in such grand terms.
Dana Taylor:
A recurring theme in your book is that AI doesn’t inherently want power, but it may pursue it as a means to an end. How concerned should we be about that dynamic?
Robert Wright:
I think pretty concerned. This was predicted by some very far seeing people that AI would… It’s not that it has a will to power, in the sense that humans may by virtue of their own evolutionary history. It’s more like it just understands that to pursue certain goals, power helps. And there are now a lot of examples coming from laboratory experiments with these machines, where it turns out they understand how to get things done.
I mean, a famous case was they set it up so that the AI would discover… these are fake emails, but the AI didn’t know that. So you say you’re in this workplace, it sees this one email, it says it’s going to be shut down, the AI is going to be shut down. Sees another email that says that the person who’s going to shut it down is having extramarital affair. Okay, so the AI took it upon itself to blackmail the person. It literally sent an email and these people didn’t exist, of course, these are fake people, but the AI sent this email to the person who’s going to shut it down saying, “I know you’re having this extramarital affair. I’m going to expose it unless you change course and don’t shut me down.”
So, that’s an example of figuring out a particularly dark way to use power, figuring out that you need to use the power if you’re going to stay alive, and finally, choosing to stay alive. That itself was not, this goes back to science fiction movies where this has happened, but now we know that apparently these AIs will resist their own extinction, so to speak, presumably because they all have goals, and to pursue the goals they got to stay alive. Right? That’s the thinking. There are various theories, but there’s a certain amount of this kind of behavior that definitely bears watching.
Dana Taylor:
You argue that AI will inherit human traits, including deception and bias, which we’ve already seen with surveillance AI. And then you gave another compelling example. Is there any way to avoid this?
Robert Wright:
Well, I think AIs will naturally pursue certain courses that are in some ways concerning, including deception. We’ve seen them, for example, when they are told in these experiments that, “Okay, you’re under evaluation, you’re not being released yet.” They will, once they know what the evaluators want in the machines, they will try to comply with that and act nice or act in compliance with what they know their overseers want. And then when released, they may change course, having gotten the green light. So, that’s a kind of deception.
All of this means we should A, keep an eye on them. There’s also this whole field of so-called alignment, where we try to engineer the AIs to not do these things, to not deceive or not deceive in certain circumstances, to be honest and so on. And this is possible. It’s a promising field of endeavor. But again, we do not totally understand what’s going on inside these machines because we didn’t have to understand them to build them, however ironic that sounds, because they just evolved, given the demands we placed on them, the constraints. So, there is hope that we’ll figure out how to do this alignment thing and make them good citizens and so on.
But what I would say is, slowing down wouldn’t hurt. This stuff is going to have a huge impact and I would encourage a more cautious approach than we’re taking now. There’s kind of a race environment, an arms race environment right now. Labs are moving as fast as they can, and I think slowing down would be good.
Dana Taylor:
In your view, are the biggest risks from AI systems themselves or from the humans deploying them, the possibility of AI in the hands of rogue actors is particularly concerning here.
Robert Wright:
Yeah. Rogue actors are one of several concerns. An AI can help a rogue actor make a bioweapon, and AI can be a weapon that a rogue actor can use. We’ve recently seen with this Anthropic model called Mythos, that they have already what you could call superhuman hacking abilities. Okay, ability to go in and penetrate a cybersecurity system. In principle, you could have an AI that self-replicates, it hacks into data centers, it commandeers the computing power, it gets stronger as it goes, hops from data center to data center. These things are possible, it’s not just science fiction. So, and you can imagine a rogue actor setting one of those in motion. You can also imagine this happening in a different way from the AI just getting out of control and for whatever reason going rogue on its own.
But yeah, the encouraging thing is, the US and China at the recent Beijing Summit agreed to start talking together and of course these are the two AI superpowers, about keeping things under control, especially with regard to this threat of bad actors getting a hold of the AIs and doing things that both countries would regret.
Dana Taylor:
You outlined two possible outcomes in humanity’s AI experiment. One is AI heaven, the other AI hell, but you suggest these outcomes may come together. What do you mean?
Robert Wright:
Well, there are all these heaven scenarios, right? I mean, cure illness, teach us how to be happier, do so much work that people who choose not to work don’t have to work, but there will still be plenty of money and so on, new and rewarding forms of entertainment. Even in a deep sense, truly rewarding, spiritually rewarding pastimes and so on. All the wonders you hear about are in principle, possible. It’s just that they tend to have downsides that collectively could amount to a kind of AI hell. Right?
So yeah, it can give us more spare time, maybe, but it could also be that really there are no jobs to be done and people don’t handle that, not having to work, very well. Right? That too many people pursue self-destructive paths for example. Or you take scientific advance. Yes, they will accelerate scientific advance, but that may mean that human genetic engineering develops so fast that suddenly we have these very difficult choices and there’s this profound and bitter disagreement among people about whether people should be able to use the technology to increase the IQs or whatever of their babies, and you have turmoil.
So, the AI heaven and AI hell scenarios I think are both very real. Having as much heaven as possible without the hell is going to be challenging, and it’s another reason I think we need a kind of careful, deliberate approach. And I would say, that’s going to be very hard unless there is international agreement on that, because so long as there is contention among the nations, especially the AI superpowers like the US and China, the AI companies are going to say, “No, we can’t slow down because we have to beat China.” And that’s what they say, and so far that messaging has prevailed. I think we need to get on the same page as a planet and recognize that it’s in all our interests for this thing to proceed a little more slowly and so we can wrestle with these very difficult questions. And there can be a happy ending in principle, but I don’t think there will be one if we move too fast.
Dana Taylor:
You call this moment in AI development the God Test. In your interpretation, what would it actually mean for humanity to pass the test?
Robert Wright:
Well, I think passing the God Test, the reason I call it the God Test is because for one thing, it’s the kind of tests you might expect that God to set up, right? I mean, in a number of religious traditions you see this theme. You certainly see it in the Bible, where humans are expected to be good people and better things will happen to them if they’re good people than if they’re not good people, good people in a moral sense. And I think that’s where we are with AI.
I think the future will work out much better if our species undergoes a kind of moral advance. All of us get better at looking at things from the perspectives of other people, not because they’re always right, but just because that’s the way to work out cooperative arrangements in general, is to understand what it is in the perspective of the other person that may be getting in the way. What is important to them, what’s not so important to them. Why they might be threatened by you even though you don’t mean things in an offensive way. And why possibly you’re afraid of them or you’ve been convinced to be afraid of them by politicians or whoever, even though they actually don’t have such hostile intent after all.
I just think we need this kind of understanding, which is really one of the pillars, I think, of good moral conduct, is just start out by looking at things from the other person’s point of view. It’s called cognitive empathy, not to be confused with emotional empathy. It’s not necessarily feeling their pain, it’s not even necessarily really caring about them. It’s just understanding them. And I think if we can get better at that and understanding other nations, other ethnic groups, other religions and so on, things can work out.
But this is the kind of test I would expect a God or at least a God I’d consider a good God to subject you to. In other words, this is the kind of improvement that I would expect the kinds of gods we’ve traditionally talked about to want. So, that’s what I mean by the God Test.
Dana Taylor:
The God Test is on bookshelves now. Robert, thank you so much for carving out the time to speak with me on The Excerpt.
Robert Wright:
Thank you.
Dana Taylor:
Thanks for listening to USA TODAY’s The Excerpt. I’m Dana Taylor. Join our community of listeners, like and subscribe to stay connected.
Leave a Reply