Here, he says:
"I'm very interested in historical aspects of [AI] research. …we keep finding research papers, peer-reviewed papers, going back to the '80s, '70s, '60s--I think the earliest is 1885--where people explain this exact problem: we'll have machines; they'll become better than us; we'll have competition. I was curious--well, how far back does it go? ...AI has history in philosophy, in theology, early philosophers. If you look at theology--forget about whether religion is true or false, just as a study--people took a number of assumptions: there's a creator, he gave us a book of instructions. From that point on, you can see there's really a struggle of creator to control creation. And all the concepts, if you give them scientific names, map on [or parallel] perfectly. You have this designer of biological robots [humans] who wants to give them an ethical code, rules of conduct, reward and punishment--everything maps on perfectly! And the fact that after, I don't know, 5000 years of research, theology didn't come to a solution is not very encouraging that we are going to [have one] in the next 30 years. It seems like a huge problem and what we're doing is actually harder. In theology you have superintelligence trying to control lower intelligence. We are trying to control god. We are trying to control a superintelligent being."If you decide to listen to the podcast, he says this about 49 or 50 minutes into the interview. He is then asked whether this is just a coincidence or whether there's is something deeper going on. Roman replies by referencing the book, "Superintelligence," where Nick Bostrom explains that we could be living in a simulation, implying that a superintelligent entity or designer has created the simulation. To someone that doesn't believe in God, this is probably the closest admission you'll get to a creator/intelligent designer. An atheist might say, "If given a choice, I would be much more willing to accept that the universe is a simulation by some kind of artificial superintelligence vs. the biblical account of creation."
Then there's another view. Recently, Britain's Astronomer Royal, Lord Martin Rees, who is also a member of SETI (the Search for Extra-Terrestial Intelligence), wrote an op-ed in the Financial Times explaining that if we are to find extra-terrestial life somewhere in the universe, it is quite likely it will robotic or some form of artificial intelligence. At the very end, he says, "They may be our own remote descendants."
So, there you have it. Don't believe in God or the bible or any of that narrow-minded stuff? No problem. At least you can believe in the next best thing: an artificial superintelligence created us--either by simulation or by direct decent. And if it shows up and says, "Hey there! I created you and I'm God. Worship me cause I'll help you achieve immortality and all that stuff that people have been looking for all these years"--what will all the atheists, non-believing scientists, and leaders of the world do? If there's a story they're willing to believe, this is it. They're already telling us what to expect.