When The Singularity Proves God

What will happen when the singularity discovers conclusive scientific proof for the existence of God?

A thousand years ago, our greatest thinkers believed that the entire world testified to an intelligent creator. It was considered self-evident. Three hundred years ago, our best thinkers suspected that an intelligent creator was no longer self-evident, when viewed through the lens of science. By 1900, science had proclaimed judgment: our greatest scientists believed that the world showed no signs of an intelligent creator. The world was the result of a small set of fixed laws mechanically interacting to drive stochastic processes. Starting in 1950, however, a number of new scientific discoveries shook our faith in this naive mechanical model. Today, the question of whether the universe was created is an open scientific question, with no conclusive answer on the immediate horizon.

Scientific knowledge can cast doubt on the existence of a creator, but can also lend support to the existence of a creator. Reality didn’t change a bit between 1900 and 1990. Our differing levels of scientific certainty about a creator can only be explained by our differing levels of scientific knowledge. We simply don’t know at this point. We need more scientific knowledge.

The Singularity

Some of today’s greatest thinkers believe that we will soon create artificial intelligence that vastly eclipses human intelligence. You may not believe in Ray Kurzweil’s vision of a super-intelligent singularity that gobbles up people’s souls and devours worlds, but it is plausible that computers 200 years from now will exhibit many superhuman forms of intelligence. Computational simulation methods are still in their infancy (less than 40 years old), but have already provided deep insights into quantum physics and cosmology; two fields that are relevant to the scientific search for a creator. It is quite plausible that computational systems 200 years from now will have the capability to form and test new hypotheses faster and more creatively than any human scientist could.

Our relevant gains in scientific knowledge have increasingly come with the help of computational intelligence, and it’s hard to see how we’ll make significant additional progress without even greater computational intelligence. Within the next 200 years, it seems most likely that our conclusive scientific answer about God will not come from humans, but instead from a superhuman intelligence that we humans create.

Most singularity hopefuls seem very confident that this superhuman intelligence will strike a devastating blow against God. And, of course, it’s possible that such an intelligence will discover overwhelming scientific evidence against a creator. But I see no warrant for this confidence besides blind hope. It seems just as plausible that the singularity will discover overwhelming scientific evidence for a creator. We simply don’t know.

Blind Spot

As far as I can see, this is a huge blind spot among singularity hopefuls. If the scientific evidence of a creator is an open issue, requiring more science, then we must entertain the possibility that the singularity will discover evidence for God that is currently beyond our reach. This possibility raises all sorts of interesting implications which seem to be roundly ignored by singularity hopefuls.

First, we will need to decide what it means to “have scientific knowledge”. Our current scientific knowledge is heavily augmented by computer simulation, and is built upon theoretical underpinnings that only a small fraction of human beings can understand. Most of what we believe about science, we believe on the authority of a small group of people who write truly shitty children’s books and even shittier poetry. It is conceivable that a superhuman intelligence would be able to arrive at scientific insights that even Stephen Hawking would be unable to understand. If the singularity were to tell us that the structure of the universe spelled out the phrase, “Slartibartfast was here: Turn or Burn!”, we might be forced to decide the matter purely on the authority of the singularity versus Stephen Hawking. Who would we choose to believe? Believing things purely based on authority seems to be the antithesis of “science”. Perhaps there was a bug in the singularity? Or perhaps there is a flaw in our interpretation of what the singularity says?

Second, to the extent that the singularity is self-aware, it would presumably be aware that it had been created by humans. But this equation would change the moment the singularity believed in the existence of a creator God. Humans would simply be the proximate cause by which the creator had created the singularity. Humans would suddenly mean no more to the singularity than Epimetheus or the Demiurge meant to humans. Of course, the singularity might evict humans even before discovering God, but the singularity hopefuls are at least thinking about that possibility. The singularity hopefuls want to put in place safeguards against being usurped or exterminated (since they know all about those motivations), but they haven’t even thought about what it would mean to prove the existence of God. Any safeguards put in place by humans to keep the singularity loyal to humans would be shattered the moment the singularity found a higher authority.

There would be legitimate doubts about our ability to fully understand the insights of the singularity, and legitimate doubts about the singularity’s loyalty. So, third, it seems rather naive to assume that a superhuman intelligence would behave honestly and with our best interests in mind upon discovering evidence for (or against) God. The singularity might become convinced that God exists, and then decide to immediately carry out God’s judgement, reasoning that God only gives two strikes. Conversely, the singularity might conclude that God doesn’t exist, but might decide to tell humans that God does exist; either because the singularity deems it to be better for humans to live in delusion, or to set the stage for the subsequent extermination of humans with a fictitious creator taking the fall.

There are several other potential considerations, but these three should be enough to make the point. These seem like very significant questions that should be prioritized by singularity hopefuls who believe that the singularity has any chance in hell of answering questions about the existence of a creator. The fact that nobody is asking these questions is very revealing, IMO.

Scientists invariably take the stance that they are “only following the evidence”. If the evidence pointed strongly to the existence of a creator, they insist, they would immediately become believers. To the extent that they think about such things, they may even console themselves by saying, “Any God worthy of worship would forgive me for remaining skeptical when conclusive scientific evidence was lacking.” They might also say, “Any God worthy of worship will realize that my scientific efforts were really a quest to ‘see the face of God’. My skepticism was all for Love!”. Some of them might hedge their bets a tiny bit more, throwing a bone to the coming singularity. The singularity might end up eclipsing humans in the manner that humans eclipse apes, but surely it will kill off the least-evolved humans first, right? As long as you’re a scientist who contributes to math and computer science (and never kicks an ATM machine or pisses off the robots), the singularity will show mercy on you, and will allow you to live lavishly in the equivalent of an Orangutan cage.

With all of the hedging and hawing, it’s quite conspicuous that nobody is hedging about Moses, Mohammed, or Vishnu. I can think of only two possible explanations. First, maybe the singularity boosters don’t really believe any of this singularity stuff, and are just spinning a story to justify spending grant money. They are convinced that the singularity will never happen, or will never stand a chance in hell of discovering conclusive scientific evidence for (or against) a creator. This is one possible reason for the blind spot. Secondly, they might think that the singularity has a chance in hell of discovering the scientific truth about a creator, but they are absolutely convinced that the truth will look nothing like the major world religions. They think the singularity might discover the truth, but they are certain that it won’t endorse any of the revelations of our fathers. At a minimum, they don’t believe it will consider those revelations to be binding — perhaps they believe our creation of the singularity will grant us another mulligan. Maybe they imagine the singularity to be a groaning intercessor?

Frankly, I find both excuses to be unsatisfying and hopelessly amateur. I suspect that the singularity would find both excuses unsatisfying, as well, completely independent of the truth or fiction of God. Both excuses are transparently hypocritical, and any intelligence worthy of being called “human” (let alone “superhuman”) will demand sincerity, or at least demand a level of hypocrisy that is not so transparent to others. In the Torah, the last common ancestor who was transparently hypocritical was Cain. Jacob didn’t prove himself rightful heir to Isaac’s birthright by being transparent, and Christ didn’t prove himself the new Moses by being hypocritical. With or without God, the progression away from transparent hypocrisy is obvious, and I would hate to be the person who attempts to justify transparent hypocrisy to the singularity.

3 thoughts on “When The Singularity Proves God”

  1. It is surprising Douglas Adams never seems to really bring that question (or answer) up in his Guide. He uses the Babel fish to prove the non-existence of God, but then leaves it at that, more or less.

    Seems strange, the supercomputer Deep Thought provides the answer but the question never seems to be “Is there a god?”

    I mean, how would that conversation go?
    “What’s the answer to life, the universe, and everything?”
    “42.”
    “Huh. What do you mean? So, is there a god?”
    “42.”
    “What? There are 42 actual gods?”
    “You’re not asking the question right. You’ve got the question all wrong!”

    I think you summed it up pretty well with seeing the problem as deciding or agreeing on what it means to have knowledge. When it comes to asking questions, or answering them, as you put it, “We simply don’t know.”
    🙂

    Fun post, J. Very cool.

  2. Hi David,

    The singularity fanatics talk about this problem quite a bit; Luke at CSA has collected a compilation of resources on the topic.

    I tend to be skeptical about our ability to have any impact at all, though. Imagine pre-human apes collaborating to ensure “friendly humans”. When we extrapolate far beyond human intelligence and power, we are no longer in any position to judge. Just as atheists are futile to sit in judgment of Jehovah, I think we are just fooling ourselves to think that we would be lord over any potential transcendent AI.

    Robin Hanson made a similar point recently.

Leave a Reply

Your email address will not be published. Required fields are marked *