AI & Religion

Ego_SpeakPeople who do religion don’t need to study it, because they already know what it is. People who have never embraced a religion as an adult feel they can study religion objectively and still really understand it. Religious folks feel they can fairly evaluate the “truth” of other religions by reading books about them. People who have butterflied around several faiths over the years but never made a long home in one, feel they understand what drives the religious heart. Grand generalizations about religion spew easily from many of these people. But are the generalizations intelligent?

brainWhat can sneak by as “intelligent” is often over-rated. Since my youth, I’ve watched normal conversations and seen superficiality constantly disguising itself as depth. One summer, in my late teens, I read the book “EgoSpeak” which confirmed my worst skepticism. Perhaps God can rescue us from our meaningless chatter, I thought, as I dove into Christianity at that time. You see, my god, my religion and much more were colored by my temperament and my circumstance. But it didn’t take too long, as I watched my own conversations with God, until I saw the same EgoSpeak creature doing the same things.

Intelligence that can step back and understand circumstance and context is hard to culture. Most of our conversations are on a locomotive of habit — a juggernaut of trite conversation grabbers full of tricks to pretend we are listening to the other.

The New Yorker has a fun, short article reviewing a paper which criticizes the type of artificial intelligence which aims at mimicking human conversation (consider “Siri”). I agree with their conclusions that creating tricks, diversions and ploys to enable a machine to sound like an actual human interaction is failing because we are only building machines which avoid a whole class of conversations — insightful one. But I worry that the criticism can go deeper. Is normal human speech (“Ego Speak”) really something worth emulating — even when it is “insightful” or intelligent?


NPR and the New York Times recently educated me on something called “social bots” — AI which fool readers into thinking they are interacting with a real person.  Apparently, they flood on-line dating services and trick people constantly.  One entrepreneur bought such a service, and went into the data, finding these bots and erased them.  The surprising result of cleaning up the service: membership dropped!  People loved being told how cute they are, how interesting they are, how much someone adores them — even if those people keep disappearing or seeming strange after more indepth conversation.  The bots keep real people coming back for more superficiality.

Soccer-RobotThis month a big stress dropped from my life when I again passed my medical boards (every 6 years).  So, as you may have noticed, I have more time to write now.  Taking advantage of this time, I am also taking some computer courses with my son at  And as I dive back into programming, and seeing this program on youtube which uses neural and genetic programming to generate try and mimic real huma soccer playing, I daydream of writing modeling software (a delusional ambition) which mimics human religious behavior. But would such a project be as disillusioning as the reading the book EgoSpeak, as learning about the deceptive power of Social Bots and as watching people theorize about religion.  I think so.

Generating grand theories based on our own limited experience seems similar to EgoSpeak where we grab conversations back into our own court. We rarely deeply encounter another person and yet feel we encounter people all day long. We can’t see ourselves clearly. Is this human stupidity worth emulating on computers? What would a higher intelligence look like — perhaps one day machines will show us. Humans have feigned intelligence for a million years and created imitators to fool each other: both gods and bots. I’d say we shoot higher than imitating human.


Image Credits: Brain, Titter-bot, Soccer-bot (btw, can you tell if this is bot generated?)


Filed under Philosophy & Religion

6 responses to “AI & Religion

  1. Ian

    Glad to hear you’ve got some more time to write, congrats on the medical boards.

    So I got a little lost in the first bit, sorry. It seemed a little mired in futility. I think we can be insightful and intelligent and discover general principles. And I think the traps that trick us into thinking we’re being profound when we aren’t can be detected and avoided. So I didn’t agree with the first bit.

    On dating AI. Interesting. I’ve worked on dating AI, and it is *definitely* true that the best matching doesn’t give the best client experience. This is an instance of a key point I’ve come across, which I’ll blog about this afternoon rather than discuss here. I haven’t worked on compliment bots, however, but they are a common bane of social services (not just dating, Facebook, Twitter, etc). In my experience these are not always AI though, interestingly they are often “Mechanical Turks” – very low paid offshore labour following a very specific script. Or a program running on a person’s hardware.

    The really interesting thing is the Religious AI. I do believe that there comes a point where something is such an accurate mimic of something that it makes no sense not to call it that. So Turing was right, an AI sufficiently advanced to be indistinguishable from human would be intelligent, I think. Or, a religious AI whose behavior is indistinguishable from a human believer is religious. BUT! The *very* important question, unasked, is about non-human intelligence. Okay, if we can simulate a person indistinguishably we have a person. But that’s likely to be *far* more difficult and complex than allowing the AI to be intelligent in its own way. We can’t yet make a running bipedal robot that moves as fast as a sprinter, so have we failed to make a mobile machine? No, of course not, we can make things that go far faster than people can. I think the same thing is true of intelligence, and, by extension, other patterns of behavior. So the interesting question to me is not, how do you make a computer appear to have human religion, but what would an AI that is religious in its own terms look like?

  2. @ Ian,
    Yeah, this morning I debated writing poetry or just rambling about stuff I was thinking about. You can see the overlap — but it was fun inconsistency.
    Thanks for contributing.
    I agree that progress in knowledge is possible — obviously so.
    But in some arenas we fool ourselves more than others.
    I also don’t think that all the tricks that trip up our think can be caught.
    But I am in favor of building safety guards and we have made wonderful tools which are continually being developed.

    I am looking forward to you post on dating.

    Allowing AI to be intelligent in its own way, was part of my point and you make it much better (as usual).

  3. TWF

    Interesting stuff. There are degrees of depth and understanding which we achieve. I can see Ian‘s point, but also Sabio‘s counter, and would suggest that you’re both right, because there is a spectrum of moods and capabilities. Yet I think that Sabio‘s view represents the “lowest energy state”, and thus is likely most common.

    I like the thought of driving AI in other ways than humanity-based likenesses, but, on the other hand, I can see a use for such mimicry. Even if we never quite get to insightful conversations with robots, the beauty of the superficial communication is manifold:

    1) We don’t generally seek out insightful conversations with everyone we meet, like at the fast food counter, so such effort would be generally wasted.
    2) We generally like the superficial lies better than insightful truths, and these lies serve our mental well being to some extent. (As social bots show us, and how we intuitively know to pay a complement instead of reply honestly in certain situations.)
    3) Enabling this superficial AI makes human-to-machine interface generally much more ergonomic and intuitive, not to mention how this superficial speech-driven interface aids in technological accessibility for a certain portion of the disabled populous.

    So I say let’s continue down both paths; the deep and the shallow! Swimming is great, but so is splashing in the puddles!

  4. @ TWF:
    Don’t you ever side with Ian again. Common man, we are both Americans — where is your tribal ethics — where do you find “Identity”? Laughing. This is a segue to my coming post! (but your comment has distracted me, and I shall probably write something else today.

    I loved your “lowest energy state” analogy and could see two sorts of graphs pop up in my mind as read your words. In nuclear physics, we learn of electron orbital energies (when we are not polluting our minds with chatter of probability) – and there we imagine “lower” states. But in chemical kinetics, I could think of two sets of graphs, one with a “lower energy state” determined by the second law of thermodynamics, and the other a more complex reaction involving stable wells of energy levels with lower levels of energy below but barriers to hinder the “fall”. Wow, not that I think of this, I will have to do some graphs to illustrate this in relationship to religion.

    Your point about a the value of small talk is, of course, important.

    I agree: working on “both paths” is important.

  5. TWF

    Indeed, I had electron orbital theory in mind with that comment. 🙂 But chemical kinetics works just as well too! Though maybe not as discrete as orbitals, I know my mind works at different energy “states”, with the most analytical and focused, and particularly the view of abstraction to where I can see my own actions objectively, being at much “higher” states than where I am most of the time.

  6. rautakyy

    Worth to read is the Stanislaw Lem book Kyberias. A nother good read is Isaac Asimovs The Robots of Dawn. Both deal in interresting suggestions about the possibilities of AI.

    As one of those who has never been part of any religion and having allways had the outsiders look at them, I could say, yes, that I think on certain level I understand them better, than the adherents, most of whom have never investigated their faith and the few, who have investigated it whith the sole purpose of justifying what they allready believe. I do not think this is so much a question of intelligence as it is of information. Which in turn gives tools for the mind to figure out what is true. To me at least religions present themselves as easy answers to difficult questions, ladden only whith superstitious evidence.

    What I find curious about our relationship to AI is how easily we the audience accept the status of robots in the Star Wars movies. Yes, those are just space opera entertainment, but even so, have you noticed that the robots expressing what seems as genuine humane emotions are treated as property by both the baddies and the good guys. Even if they are just simulations of human emotions, what should we make of people who so openly dismiss those rather convincing expressions of emotion? Now, it is possible that in “a galaxy far, far away, long time ago” their culture percieved these machines whith very advanced AI as property, but how is it we modern people see this form of slavery as a-ok, or do not even notice it?

Please share your opinions!

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s