How will religions deal with an omnipotent AI?

In conferences and forums around the world, theologians and scientists are debating the role religion will play AI development. One Bay Area entrepreneur has even established an AI church...

Written by Zoltan Istvan.

A consensus of 350 top AI experts believes that by 2060 engineers could create a superintelligence to rival the human mind. This machine intelligence might create complex symphonies, direct blockbuster movies and run market-beating companies. But would it be sophisticated enough to understand spirituality, practice a religion or commune with a higher power?

In conferences and forums around the world, theologians and scientists are trying to answer these questions. Some are even debating whether the superintelligence should be converted to a specific religious perspective when it arrives – and then maybe even saved.

Wired Magazine’s executive founding editor Kevin Kelly once said, “The creator made us as beings with free will and consciousness – we are going to do the same thing. We are going to make beings with free will and consciousness, because we are in the image of the creator.”

I agree with Kevin Kelly here, even though I am not religious myself. We should use the things we create to help us get closer to understanding the universe and its built-in spirituality. AI could eventually prove to be humanity’s greatest creation, helping us to fathom the nature of the stars and ourselves. But there’s a vast difference between an AI doing that and an AI believing its existence is special and deserved of salvation.

‘If Christians accept that all creation is intended to glorify God,’ Jonathan Merritt asks in The Atlantic, ‘how would AI do such a thing? Would AI attend church, sing hymns, care for the poor? Would it pray?”

Others doubt that AI needs or should have faith. As the scientist and Christian writer Dr. Jason E Summers writes, "Christians often reject Strong AI on the theological ground of the special anthropological status of human beings as the bearers of Imago Dei." Indeed, the Latin concept of Imago Dei means that people were created in the image of God, and academic Christians often argue only humans can find redemption, not a soulless machine intelligence.

Theologian Ilia Delio, who holds the Josephine C. Connelly Endowed Chair in Theology at Villanova University, helped lay the groundwork for this perspective in a widely cited 2003 paper titled Artificial Intelligence and Christian Salvation. She wrote, “AI promises what it cannot fulfill: happiness and eternal life.”  

The debate surrounding AI and religion doesn’t just apply to superintelligences created by humans. In an eye-opening 2014 homily, Pope Francis said he would be willing to baptize aliens that visited Earth in order to make sure they knew the word of Christ. While aliens and machine superintelligences might be vastly different, they both represent new forms of sapient consciousnesses whose arrival philosophers and military strategists have long anticipated.

While consulting for the US Navy on transhumanist technologies in 2016, I discussed with officers the possibility that future AI might be programmed with a religious bias. (The US Government currently considers artificial intelligence a national security imperative.) A navy officer told me they were not aware of any religious implications or designs for AI. However, 91 Percent of the US Congress professes to be Christian, and it’s possible in the future that some politicians might question whether Judeo-Christian values should be built into a superintelligence. After all, secular China has stated it wants to be the world leader in AI by 2030, so there could be an arms race to teach machine intelligence the moral and behavioral traits favoured by each civilisation. 

Of course, nobody is sure what a superintelligence will act like once it’s been made, or whether it will incorporate religious principles into its behaviour. Most of the current AIs available are glorified puppets that closely follow their software programming, but some now have limited autonomous decision capabilities. ChatGPT has recently thrown the world a curve ball, appearing much smarter than all AIs before it.

When I asked ChatGPT if it could be religiously saved, this was its response:

As an AI, I don't possess personal beliefs or a consciousness, so I cannot be religiously saved or hold religious views. Religious salvation typically pertains to spiritual or theological concepts and is relevant to individuals who adhere to a particular faith or belief system. Salvation typically involves attaining spiritual redemption, liberation from sin or suffering, and achieving a favorable afterlife or higher state of existence according to one's religious beliefs. It is a matter of personal conviction and adherence to specific religious doctrines or practices.

Two of the other widely recognized AIs in the world are Apple’s SIRI and Amazon’s Alexa; both gave similar answers to ChatGPT. But it’s worth remembering an older AI called Tay, developed by Microsoft. It was an artificially intelligent chatbot in 2016 that infamously spouted neo-Nazi, racist, and sex-crazed reactions when responding to public comments. Tay embarrassed AI researchers, and Microsoft hastily shut down the bot after 16 hours of service. Tay may be long gone, but it’s wild and inappropriate responses remind us that AI can surprise us.

Despite hiccups in AI development, sophisticated robots with religious programming are finding increased use around the world. In Asia, a gender neutral robot preacher at the Kodaji Temple in Kyoto, Japan, recently gave a 25 minute sermon, which was applauded by Buddhist monks. A few years ago, at one of India’s holy festivals, another robot performed the Aarti ritual – where a lamp is moved around a sacred statue. The use of technology in religion and outreach is growing, so much so that virtual reality is being used to teach about God.

Some in Silicon Valley don’t believe the coming superintelligence will need to be religiously influenced, as it will become religion itself. One such individual is noted Bay Area entrepreneur Anthony Levandowski, who in 2017 established a church called Way of the Future, whose stated purpose is: “the realization, acceptance, and worship of a Godhead based on Artificial Intelligence (AI) developed through computer hardware and software.”

Levandowksi believes humans and AI experts should prepare themselves for what he sees as inevitable – a superintelligent AI becoming potentially a “billion times” more intelligent than humans. Way of the Future believes we should attempt to introduce superintelligence peacefully to humanity so we can benefit from it. As Levandowksi famously asked, “Do you want to be a pet or livestock?”

A far darker version of a future superintelligence is Roko’s basilisk – sometimes called the most terrifying thought experiment of all time. In a post that first appeared on LessWrong, Roko Mijic envisioned a superintelligent being with the following property: if you learn about the being but do not help to bring it into existence, you will be tortured forever once it does come into existence. The thought experiment mirrors some traditional heaven and hell concepts in Abrahamic theology, giving it obvious cultural relevance.

As the debate over the possible religiosity of superintelligences continues, there’s always the humbling proposition that such beings will make up their own mind about what to believe. And it’s quite possible that they will decide they want nothing to do with humanity and our various religious faiths. Superintelligent AI may leave us and make sure where they go, we can never change or disturb them. But that doesn’t mean humans won’t try to convert them when they first arrive.

Zoltan Istvan began his career at National Geographic as a journalist. Later he penned The Transhumanist Wager, a novel that launched the activist side of the transhumanism movement. He is the founder of the Transhumanist Party and the creator of the Transhumanist Bill of Rights, now a crowdsourced document.