Blog: AI as the asymptotic religion
A contrarian take on how AI could shape the trajectory of humanity. And how the apparent polarities of religion and cutting-edge technology are perhaps two ends of a necklace joining to form the asymptotic whole.
I walked into a room of academicians just as an experienced AI researcher looked up from his mobile and commented ‘Someone is talking of using AI for spreading religion.’ All laughed. Someone else added, ‘Perhaps, AI is the new religion.’ Some smiled. But it got me thinking. After struggling with the idea for a while, here are some thoughts.
Many articles on AI do evoke a sense of mystery, doubt, and even fear amongst the general public. Some of these emotions are not misplaced. With AI as a technology, we have accomplished things hard to imagine a decade back. And reasonable extrapolations of this rapid rise do suggest job losses, an AI arms race, concentration of power, and rise of AI super-lords. I would in this piece pose a contrarian opinion while not contradicting the above. I.e., while I see all of the above challenges as realistic, I see them being counter-weighted by a potentially strong super-outcome.
Before we figure out what that super-outcome is, let us note our working definitions for two jargon minefields:
AI = ability of non-biological entities to perform complex goal-oriented tasks
Religion = a belief system of a conscious entity to construct and verify a unifying substratum to all lived experience
I concede there are twice as many holes in the above as there are opinionated people, but we will have to work with these for now.
I see us humans as having three spheres of mental activities:
1. Reasoning / decision making
2. Forging individuality
3. Experiencing reality
Often, these are related. If you newly experience joy on a trek on the Himalayas, you would work on modifying your individuality to favour solitary travels to mountains, and then rationally set aside savings to finance such travels.
More momentarily: When playing chess there is the rational decision making of the 1,729 possible moves for a particular piece, then there is the style of attack that you define as your individual playing style, and finally there is the experience of being checkmated.
So three incessant activities, going on together.
I see the rise of AI, indeed of technology, as providing the potential to shift the balance of our mental activities.
For instance, it turns out that we humans are mostly irrational, and also irrationally naive about our irrationality. Thus automated stock market agents are usually better at managing investment portfolios. Even our brightest samples cannot match against AI agents on narrower tasks such as playing Jeopardy, Chess, Go, and DOTA. Indeed, in all spheres where we can capture holistic information systematically and have the financial motive to sponsor large compute resources to train AI agents, humans are likely to be beaten. This is the state of affairs today. As time moves, more data and more compute would make AI decision-making pervasive.
Some of us may heartily relinquish right brain number crunching and take refuge and subsequently pride in our humanising creativity and individuality. There is succour there, but with an expiry date.
As we increase the bandwidth between ourselves and the informational world, we provide more power to AI agents to decode our individuality. Indeed, based on shopping histories of myself and people like me, Amazon can suggest products which I would like but never knew existed. Same for songs on Spotify, movies on Netflix, or partners on Tinder. While these are transactions on the internet, but it does not stop there: With the rise of wearables and implants, we could expose more biological information about us than we are consciously aware. Interconnecting brains and AI agents does not quite seem a pipe dream any more. And at that point, it would be very hard to claim that I know more about myself than my AI agent.
Sure, this post-AI world is still sometime away — maybe years or even decades away. But it is likely not centuries away — the blinding fear of death will continue to motivate us as a species to innovate in biotechnology and other areas and get there as soon as possible.
In this post-AI situation, most of us are on shaky ground. Machines are doing the physical work. AI agents are doing the decision making and forging our individuality. What do we then do? What defines you? Who am I?
My philosophy, in essence, is the concept of man as a heroic being, with his own happiness as the moral purpose of his life, with productive achievement as his noblest activity, and reason as his only absolute. — Ayn Rand
Descartes had said “I think, therefore I am” and Ayn Rand said the above roughly meaning “I think, I define my morality, and I do my actions, therefore I am.” But neither of these seem to be defining I-am-ness in the post-AI age.
This is the act where our other protagonist enters. Enter Religion with a quote from the Bible.
’ehyeh ’ăšer ’ehyeh — Exodus 3.14
Most commonly accepted translation: “I am that I am”.
Here is a unique take on I-am-ness, that of experience. Here is what it means: Even after I let go of reasoning and decision making and of forging my individuality, I still experience reality, and in that I am. I will hazard one more esoteric step here: The Vedas expand even beyond the above and proclaim that the experience of I-am-ness is all there is.
तत्त्वमसि (tat tvam asi) — Chandogya Upanishad
Translation: “You are That” (That refers to ब्रह्मन् — Brahman)
The proclamations of mystics, in religion after religion, is that our true home is in this bare experience. The Vedantic phrase characterises this true home as the equivalence between the three terms सत् — is-ness, चित् — experience, and आनन्द — bliss.
And so here is the crux of it all. I believe AI is accelerating the potential for an evolution of shifting our mental activities. It is providing the potential of freeing us from the logical and the ego-centered mental activities that we find ourselves bound to, and to focus on bare experience.
What more is there to be said? Well there is the matter of Darwin. The efficient consequentialism of natural selection should not be ignored. While AI provides us the potential for the above, it does not assure it. If we so choose, things can turn out differently. For instance, if we choose to relinquish how to navigate in the real world and instead focus on navigating Mario to collect virtual coins, things can turn out differently.
The way things turn out will of course depend on which choices lead to individuals who are fitter to lead lives on Earth and other planets we may inhabit. And while AI is an asymptotic precursor to religion, we as a species may go off a tangent, if the geometry has been so designed.