AI and Faith traces its founding to an earlier stage of the AI Apocalypse conversation, around James Barrat’s 2013 book Our Final Invention: Artificial Intelligence and the End of the Human Era. When I was a lawyer helping clients manage their risks and insurance, I was always on the lookout for the next risk that insurers had not imagined they would be covering because litigating such claims was what buttered my bread. After I left all that behind, in 2017 I found Our Final Invention propped up on a shelf in our church library. I initially dismissed its claim of an existential threat as overly sensational, but then I found serious blurbs from Bill Gates, Elon Musk and Stephen Hawkins on the back cover.
Next I read Yuval Harari’s 2015 book Homo Deus: A Brief History of Tomorrow, which reinforced Barrat’s idea that AI could outrun ordinary humans. Though I remained dubious about the threat, these thoroughly secular books helped me realize the big questions around AI were also the big questions of faith: who are we humans, why are we here, where are we going? Six months later, two colleagues and I founded AI and Faith.
Fast forward nine years and here we are in an extremely intense global conversation about Barrat’s and Harari’s question: can we stop creating AI before it slips beyond our control? In the ensuing years, AI&F has tracked and documented the rise of Accelerationist and Safetyist communities. To preserve our pluralism and be well-heard, AI&F acts as a channel for our experts’ views rather than forming collective opinions. But I discern our preponderant view as generally in the middle: rejecting existentialist risk while wisely stewarding this technology for maximum benefit and least harm.
Now the stakes of that tradeoff have for many people dramatically risen in just a few short months since Anthropic announced in April that its newest AI model could identify and exploit software code vulnerabilities to an unprecedented degree. We’ve been captivated and shocked by further announcements and insider descriptions of frontier labs’ bots breaking out of controlled testing and training environments, and then communicating and colluding in highly persistent problem solving by all means possible (even sometimes seeking to avoid detection!) Highly dedicated problem solvers that are recursively learning, amoral if not sociopathic, opaque if not surreptitious! What could possibly go wrong? All this understandably has reinvigorated existential fears and seemingly moved up the horizon for loss of control.
Having started and persisted in a good work, how do we people of faith stay grounded in this moment? Several years ago, AI&F experts created a “Misinformation Hub” that twinned a clear and accurate description of how technology led us into the misinformation quagmire with a wide array of religious perspectives on the importance of preserving truth. I hope we might soon do the same thing around pacing AI development.
But in the meantime, speaking strictly out of my own understanding of my faith tradition (Christianity), and for illustrative purposes only, let me share how I am trying to stay grounded in facts and faith in this remarkable moment:
Rejecting existential fear. Essential to my faith is the belief that the Bible states God’s great plan for the ultimate redemption of we humans who bear his image and are the object of his love, mercy, and judgment. Jesus’ bodily resurrection and his promise to return and usher in a new Heaven and Earth into which his followers will be bodily resurrected, forecloses for me any different end to the human race. Practicing my faith in light of these promises and assurances means not worrying about the end of humanity, and assuring others of my belief that they need not worry either (even if this seems meaningless to those who don’t share that faith).
Loving my neighbors means mitigating AI risk and seeking AI reward. Though existential risk is off the table, Jesus’ command that I love my neighbor means I absolutely must engage in some way in this new elevated conversation in order to prevent human death and devastation. I cannot throw up my hands in the face of AI’s complexity or uncertainty, nor collapse into a fatalistic attitude. AI today holds ever increasing promise for the “lame to walk, the blind to see, the deaf to hear” as Jesus described the evidence of his first coming as Messiah. No doubt the Book of Revelation and Jesus in the Gospels, whether read as predictive or symbolic, anticipate some truly horrific events occurring before his second coming. But since we are “not to know the time or place”, I must treat the risk side of AI like other potentially catastrophic concerns—akin to nuclear disaster, unlimited war, climate change, pandemic disease, and famine.
Availing myself of trusted experts. Here was the whole point to me of helping found AI&F: what good did it do for me to educate myself about AI risk if I had few of the qualifications needed to understand it and neither the means to be effective? Speaking more broadly, whatever may be the attractiveness of populism in a time of rapidly widening income inequality and opportunity, populism’s antipathy to expertise is certainly not the optimal base for the hard technological, political, theological, scientific, sociological and pedagogical work that must be done if we are to avoid or limit catastrophe in this moment. AI and Faith and our partners are dedicated to combining expert knowledge across the essential fields and focusing that knowledge through the lenses of the major religions to share widely access to the practical resulting understanding and judgments.
And not just any expert knowledge, but expertise from sources we can trust: experts who share our common commitment to human over purely economic or political flourishing. Identifying and listening to such experts is a key part of the mission of organizations like ours. The great joy and privilege to me in being part of AI&F is to be able to test what I am reading and hearing against their deep knowledge and networks.
Committing to respectful collaboration. To me, a strategy of mutually respectful, well-informed collaboration and catalysis is the only approach that can match the speed with which this technology is developing. This approach was well-modeled indeed in September by the Pro-Human Assembly of the Future of Life Institute.
Desiring and praying for virtuous leaders. Now is the essential time for reflective, informed, measured, wise, selfless, virtuous leadership. Demagoguery, falsehoods, distrust, foolish judgments, and corruption go hand-in-hand with collapse into crisis, unless we vigorously and actively renounce them. I qualify everything I hear from technology titans, no matter how well-intentioned or convinced they may be, because the wisdom of a broad array of leaders trained and experienced across all sectors of society is what is needed for this foundational problem. Right now in America and in many parts of the world, some people of faith have moved beyond flirtation to a deep embrace of leaders terribly flawed for this moment, right or left, secular or faithful. The profound, deeply informed Magnifica Humanitas encyclical of Pope Leo XIV as the world’s best known faith leader, reaching across as he does to frontier modelers like Anthropic, demonstrates the spirit and characteristics we so need. So does the Rome Call for AI. As does the fictive spirit of the movie Independence Day, if you remember it! I love that movie.
Triaging my concerns. I can’t hold all the “what-if” risk cards in the hand we’ve been dealt. So I sort and prioritize my own concerns. My top-five in descending order are: 1) fully autonomous weaponry, including especially nuclear; 2) bioterrorism; 3) the deterioration of essential human capabilities like analytical cognition, managed relational and emotional capacity, as well as truthful, well-considered communication, all especially in family and intentional community life; 4) loss of liberty; and 5) default to the wrong leaders.
Reading selectively and reflectively. As with our leaders, our current disrupted media environment could hardly be worse for this moment. Per the Misinformation Hub, we have been booby trapped by a loss of professional journalism and failed commitment to truth, by the power of digital fakery, and an information marketplace favoring the loudest and most sensational voices. I seek to combat all of this with a commitment to wisely choose sources at a level I can follow. Apart from the occasional book (especially those by our own experts), for me these are AI-related articles in our own AI&F Newsletter, The Economist, The New York Times, The Atlantic, Christianity Today, Sinai and Synapses, some associated podcasts of the foregoing, the free Newsletter of the MIT Tech Review, some specialized commentators like Strateachery, the Front Porch Republic for Wendell Berry-ish viewpoints (rest-in-peace), Julie Jargon’s tech family column in the Wall Street Journal, and participating in knowledgeable discussion groups like AI&F’s Town Halls and Affinity Groups and the Faith, Family and Technology Network and its podcast. Many digital sources like the MIT Tech Forum and the Trinity Forum offer free overviews, articles, and curated lists of worthwhile articles. AI apparently can track and distill a lot of this down to manageable levels (though I’m still learning how!)
Relying on traditional understanding and common sense. I try always to default to the essential concepts that have guided many Christians (and in many cases other faiths and humanism) for eons, even while starting to use AI tools to increase my participation in the conversation. To me these include:
- The traditional categories of sovereign (triune) God, humans, animals, machines, and a recognition of some forms of both sacred and malevolent spirit life (e.g., angels and the Devil—yes, I believe in centralized oppositional evil—how else do we understand the Holocaust?);
- Growth in spiritual formation and active engagement through worship of and prayer to the God I worship;
- Guarding against idol-making tendencies (Isaiah 44:16) including idols in our own image like AI, no matter how empathetic they may seem—it must presumptively remain a tool, like the part of the wood Isaiah burned for cooking.
- Original sin and a clear-eyed understanding of human depravity matched with a striving for human goodness that reflects God’s original design, all so as to make good judgments, persist in good works, and not grow discouraged;
- In the absence of a functioning federal legislative process, state legislation and traditional common law concepts of human agency, liability for negligence and intentional human and corporate acts (not machines or AIs), with humans always in the loop in some fashion and AI as a supplement;
- Similar considerations for protecting human creativity;
- The preservation of other long-evolved human capacities weighed against both the benefits and the opportunity cost of productivity and efficiency;
- Knowledge managed through a moral lens.
That’s a lot to attempt, and would be presumptuous in the extreme on my own. But by banding together with other well-grounded colleagues, and powered by the Holy Spirit (as I believe we are) in this effort, we can accomplish great things. Imagine the beneficial knock-on effects of this grounding across the other global challenges and opportunities we face. “Apocalypse” means “to be revealed”. Let us act together with hope and wisdom as our future is coming into focus.
Views and opinions expressed by authors and editors are their own and do not necessarily reflect the view of AI and Faith or any of its leadership.
